AI_dev Europe 2024: Full Schedule

June 19-20, 2024
Paris, France
View More Details & Registration
Note: The schedule is subject to change.

The Sched app allows you to build your schedule but is not a substitute for your event registration. You must be registered for AI_dev Europe to participate in the sessions. If you have not registered but would like to join us, please go to the event registration page to purchase a registration.

This schedule is automatically displayed in CEST (Central European Summer Time) UTC/GMT +2 hours. To see the schedule in your preferred timezone, please select from the drop-down menu to the right, above "Filter by Date."

IMPORTANT NOTE: Timing of sessions and room locations are subject to change.

07:30 CEST

Welcome Coffee

Wednesday June 19, 2024 07:30 - 09:00 CEST
Grande Galerie (Level -1)

Special Events / Exhibits / Breaks

07:30 CEST

Registration & Badge Pick-up

Wednesday June 19, 2024 07:30 - 17:30 CEST
Ground Floor Foyer

Special Events / Exhibits / Breaks

07:30 CEST

Coat & Bag Check

Wednesday June 19, 2024 07:30 - 19:30 CEST
Ground Floor Foyer

Special Events / Exhibits / Breaks

09:00 CEST

Keynote: Welcome & Opening Remarks - Ibrahim Haddad, Executive Director, LF AI & Data Foundation

Speakers

Ibrahim Haddad

Executive Director, LF AI & Data Foundation

Wednesday June 19, 2024 09:00 - 09:10 CEST
Theatre (Level -1)

Keynote Sessions

09:15 CEST

Keynote: Fireside Chat with Soumith Chintala, Co-Creator of PyTorch

Speakers

Soumith Chintala

Co-Creator of PyTorch

I am an Artificial Intelligence researcher, engineer and community builder.I am currently at Meta, jumping between Engineering, Research and Leadership as I find convenient. I also visit NYU as a part-time researcher.My career interests have been defined by two sets of work: AI Platforms/Ecosystems... Read More →

Wednesday June 19, 2024 09:15 - 09:45 CEST
Theatre (Level -1)

Keynote Sessions

09:50 CEST

Keynote: Sophia Yang, Head of Developer Advocates, Mistral AI

Speakers

Sophia Yang

Head of Developer Advocates, Mistral AI

Sophia Yang is the Head of Developer Advocate at Mistral AI. She is passionate about the AI community and the open-source community. She actively participates in various open-source communities and projects, and she has authored several Python open-source libraries. She holds an M.S... Read More →

Wednesday June 19, 2024 09:50 - 10:05 CEST
Theatre (Level -1)

Keynote Sessions

10:05 CEST

Keynote: Dan Lorenc, Founder & Chief Executive Officer, Chainguard, Inc.

Speakers

Dan Lorenc

Founder & Chief Executive Officer, Chainguard, Inc

Dan Lorenc is co-founder and CEO of Chainguard, a leading software supply chain security company. Dan has been working on and worrying about containers since 2015 as an engineer and manager. He started projects like Minikube, Skaffold, and Kaniko to make containers easy and fun, then... Read More →

Wednesday June 19, 2024 10:05 - 10:20 CEST
Theatre (Level -1)

Keynote Sessions

10:25 CEST

Keynotes to be Announced

Wednesday June 19, 2024 10:25 - 10:40 CEST
Theatre (Level -1)

Keynote Sessions

10:40 CEST

Coffee Break

Wednesday June 19, 2024 10:40 - 11:20 CEST
Grande Galerie (Level -1)

Special Events / Exhibits / Breaks

11:20 CEST

Building Reproducible ML Processes with an Open Source Stack - Einat Orr, lakeFS, Treeverse

Machine learning experiments consist of Data + Code + Environment. While MLFlow Projects are a great way to ensure reproducibility of Data Science code, it cannot ensure the reproducibility of the input data used by that code. In this talk, we'll go over the trifecta required for truly reproducible experiments: Code (KubeFlow and Git), Data (Minio+lakeFS) and Environment (Infrastructure-as-code). This talk will include a hands-on code demonstration of reproducing an experiment, while ensuring we use the exact same input data, code and processing environment as used by a previous run. We will demonstrate programmatic ways to tie all moving parts together: from creating commits that snapshot the input data, to tagging and traversing the history of both code and data in tandem.

Speakers

Einat Orr

lakeFS co-creator and Treeverse co-Founder, lakeFS; Treeverse

Einat Orr has 20+ years of experience building R&D organizations and leading the technology vision at multiple companies, the latest being Similarweb, that IPO in NYSE last May. Currently she serves as Co-founder and CEO of Treeverse, the company behind lakeFS, an open source platform... Read More →

Wednesday June 19, 2024 11:20 - 11:50 CEST
Jussieu (Level 3)

AI in Action

Audience Experience Level Intermediate

11:20 CEST

RAG Pipeline Evaluation - Estelle Scifo, Neo4j

The incorporation of Retrieval Augmented Generation (RAG) has shown significant promise in mitigating hallucinations and bolstering Large Language Model (LLM) applications, especially within enterprise-based data realms. Although various approaches can be adopted for the retriever layer, the fusion of Knowledge Graphs with RAG has showcased improved grounding and augmented explainability. As the use of RAG escalates, coupled with the different configuration possibilities, a necessity for a standardised evaluation process to gauge the RAG pipeline performance arises. Such assessment remains a daunting task, characterised by time, effort, and cost implications. Current trends lean towards utilising LLMs to minimise human intervention and to score the performance of RAG pipelines. This talk provides an overview of the leading LLM-based tools and frameworks for the automated assessment of RAG systems. It also showcases an application utilising Neo4j-backed RAG pipelines with open-source libraries that shed light on the practical implications of vector and graph-based searches for the retriever layer and their evaluation using the open-source RAG Automated Assessment (RAGAS) framework.

Speakers

Estelle Scifo

Machine Learning Engineer, Neo4j

Wednesday June 19, 2024 11:20 - 11:50 CEST
Saint-Victor (Level 3)

AI Quality & Security

Audience Experience Level Advanced

11:20 CEST

Applying Open Source Methods to Building and Training Large Language Models - Carol Chen, Red Hat & Mark Sturdevant, IBM

Several open models of varying size, quality, and performance have been released over the past twelve months. We are working on a project that allows contribution of skills to an existing model without the need to fully fork and fine-tune. This project will create the establishment of an upstream community built on contributing acceptance workflows for models. This exciting technology aims to make open source AI more approachable. In this session we'll explain how communities and individuals can contribute domain knowledge to models incrementally, in a unified and open way, while reducing model variations and resulting in an improved version by augmenting what’s already there.

Speakers

Carol Chen

Principal Community Architect, Red Hat

Carol Chen is a Community Architect at Red Hat, supporting several upstream communities such as Ansible and ManageIQ. She has been actively involved in open source communities while working for Jolla and Nokia previously. In addition, she also has experiences in software development/integration... Read More →

Mark Sturdevant

Software Developer, IBM

Mark Sturdevant is an open source software engineer and advocate at IBM. He has worked on a variety of open source projects and created many open source code patterns and tutorials to help developers get started with artificial intelligence, data science, and databases.

Wednesday June 19, 2024 11:20 - 11:50 CEST
Sorbonne Descartes & Lutèce (Level 5)

AI Research & Methodologies

Audience Experience Level Any

11:20 CEST

Powering Your Generative AI Workloads with AMD and Open-Source ROCm - Farshad Ghodsian, Sourced Group

In the generative AI ecosystem today, there is a strong emphasis on expensive AI hardware and proprietary CUDA implementations. While CUDA has undeniably played a crucial role in the success of generative AI, I’d like to share my experience with running generative AI workloads and applications on cost-effective AMD hardware and the open-source ROCm software stack. This alternative approach aims to provide users with greater flexibility and options, allowing them to apply their generative AI solutions across a wider range of hardware and software choices than ever before. Learn how to run your favourite open source large language and image generation models using ROCm, how far ROCm has come from previous versions and what features are currently supported, including PyTorch support, Optimum-AMD, Flash Attention 2, GPTQ and vLLM and how more affordable workstation class AMD GPUs compare to their Nvidia counterparts in terms of performance and inference speed. You will also see several demos of ROCm in action and some tips and things to watch out for when working with AMD GPUs.

Speakers

Farshad Ghodsian

Lead Consultant, Data and AI, Sourced Group

Farshad is a Lead Consultant at Sourced Group, an Amdocs company. He specializes in helping large enterprises build out Data and ML Platforms in the Cloud. He is currently leading efforts to develop an end-to-end MLOps platform running on Kubeflow and other open-sourced software... Read More →

Wednesday June 19, 2024 11:20 - 11:50 CEST
Monge (Level 3)

AI Systems & Performance

Audience Experience Level Beginner

11:20 CEST

Advancing Responsible AI: Unveiling the Software Carbon Efficiency Rating (SCER) for LLMs - Chris Xie, Futurewei & Tereze Gaile, Salesforce

In this talk, we'll present the SCER (Software Carbon Efficiency Rating) for LLMs initiative by the Green Software Foundation. We will share our work on a standardized approach to measure and improve the carbon efficiency of Large Language Models (LLMs), crucial for responsible AI development. We've conducted comprehensive analyses on LLMs' energy use and CO2 emissions, advocating for CO2e metrics over kWh to reflect the true environmental impact. This effort aims to drive the adoption of renewable energy in AI, advocating sustainability as part of the AI life-cycle rather than the after-fact. Join us to explore the SCER framework's potential to foster greener AI technologies.

Speakers

Tereze Gaile

Strategic Client Architect, Salesforce

Tereze Gaile is a Green Code & Sustainable AI advocate at Salesforce and a MuleSoft Strategic Client Architect based in London, UK.

Chris Xie

Futurewei

Chris Xie, Head of Open Source Strategy at Futurewei, is a key advocate for global open source collaboration. With past roles at Fortune 500 companies and startups, he offers a unique blend of technical and strategic business expertise. Recently, Chris has focused on open source sustainability... Read More →

Wednesday June 19, 2024 11:20 - 11:50 CEST
Saint Germain (Level 3)

Ethical AI Practices

Audience Experience Level Beginner

12:05 CEST

What's in a Vector Database? - JP Hwang, Weaviate

Vectors numerically embed meaning, and could be considered the language of AI. And as AI takes over the world, there's been a wide world of vector stores out there, from incumbent databases with added vector support, to fresh startups. But what's the big deal? Is a whole new product category required to store vectors? How does one even go about choosing a vector store, and deciding whether I need one? The truth is, a vector database is more than just about having an ability to store vectors. Features like vector indexes, hybrid searches, retrieval augmented generation, multi-modality and multi-tenancy significantly affect how data is stored, retrieved, augmented and isolated for users. Then, features like index types, quantization, tokenization, prompting, and replication significantly affect under-the-hood behavior and performance. And there's the matter of integration with AI models that can generate vectors, or use retrieved data to produce augmented, or transformed outputs. So join us in this talk for a deep dive into the inner workings of a vector database, and the key aspects that make them different to your grandma's database.

Speakers

JP Hwang

Technical Curriculum Developer, Weaviate

JP is a developer, tech educator, and communicator. He brings a combination of technical expertise, empathy, and commitment to all his endeavors, whether it’s through hands-on coding projects or engaging and informative talks. He believes that learning should be fun and empowering... Read More →

Wednesday June 19, 2024 12:05 - 12:35 CEST
Jussieu (Level 3)

AI in Action

Audience Experience Level Intermediate

12:05 CEST

Hallucination-Free LLMs: Strategies for Monitoring and Mitigation - Wojtek Kuberski, NannyML

The talk will cover why and how to monitor LLMs deployed to production. We will focus on the state-of-the-art solutions for detecting hallucinations, split into two types: 1. Uncertainty Quantification 2. LLM self-evaluation In the Uncertainty Quantification part, we will discuss algorithms to leverage token probabilities to estimate the quality of model responses. This includes simple accuracy estimation and more advanced methods for estimating Semantic Uncertainty or any classification metric. In the LLM self-evaluation part, we will cover using (potentially the same) LLM to quantify the quality of the answer. We will also cover state-of-the-art algorithms such as SelfCheckGPT and LLM-eval. You will build an intuitive understanding of the LLM monitoring methods, their strengths and weaknesses, and learn how to set up an LLM monitoring system easily.

Speakers

Wojtek Kuberski

CTO and Co-Founder, NannyML

Wojtek Kuberski is an AI professional and entrepreneur with a master's in AI from KU Leuven. He founded Prophecy Labs, a consultancy specializing in machine learning, before assuming his current role as a co-founder and CTO of NannyML. NannyML is an OSS for ML monitoring and silent... Read More →

Wednesday June 19, 2024 12:05 - 12:35 CEST
Saint-Victor (Level 3)

AI Quality & Security

Audience Experience Level Intermediate

12:05 CEST

Scaling to the Future: Training 100B+ Parameter Models on Trillions of Tokens - Hagay Lupesko, Databricks

In this talk we'll embark on a journey into the realm of ultra large language models (LLMs), as we unravel the complexities of training 100-billion-plus parameter models on an unprecedented scale. We'll delve into the intricacies of orchestrating thousands of GPUs to process trillions of tokens, discuss cutting-edge model architectures, training optimizations, hardware capabilities, and sophisticated orchestration and fault-tolerance strategies required to sustain weeks-long training runs. This talk is suitable for ML researchers, engineers, and anyone curious about the "sausage making" behind large scale LLM training.

Speakers

Hagay Lupesko

GenAI Engineering, Databricks Mosaic AI, Databricks

Hagay Lupesko is an engineering lead at Databricks, where he focuses on making generative AI training and inference efficient, fast, and accessible. Prior to Databricks, held led AI engineering at MosaicML (acquired by Databricks), Meta AI, and AWS ML. He shipped products across various... Read More →

Wednesday June 19, 2024 12:05 - 12:35 CEST
Sorbonne Descartes & Lutèce (Level 5)

AI Research & Methodologies

Audience Experience Level Intermediate

12:05 CEST

Building a Data Infrastructure for AI/ML - Keith Pijanowski, MinIO

The Open Table Formats (OTFs) designed by Netflix (Apache Iceberg), Uber (Aache Hudi), and Databricks (Delta Lake) have made it possible to build a cloud-native data infrastructure capable of supporting all the requirements of AI/ML. Such a data infrastructure can hold all data needed for all model types and scale out as capacity requirements change. This session will present a reference architecture for building an AI/ML data infrastructure and show how it supports MLOps, distributed training, and advanced data manipulating techniques made possible by OTF-based data storage. Additionally, this talk will show how such a data platform can support the special tooling needed for Generative AI.

Speakers

Keith Pijanowski

AI/ML Subject Matter Expert, MinIO

Keith is MinIO’s subject matter expert for all things AI/ML, where he researches and writes about storage requirements for AI and ML workloads. Keith has extensive experience in the software space, most recently as an Enterprise Architect on BNY Mellon’s Distribution Analytics... Read More →

Wednesday June 19, 2024 12:05 - 12:35 CEST
Monge (Level 3)

AI Systems & Performance

Audience Experience Level Intermediate

12:05 CEST

Towards a Practical Ethics of Generative AI - Geert Hofman, Howest, University College of West-Flanders

Gen AI has enormous potential and could be influencing billions of people in the future. The ethical ramifications are clear to everyone. There is however no need to cancel or pause development in AI for that reason. It is crucial that generative AI systems are transparent concerning their used sources and be evaluated on their output and critical capacity. This leads to the following needs: 1. Labeling systems on the openness of the sources used to generate the output. 2. Openness on the effects on externalities (human, economic and natural capitals). 3. The development of standard tests that can be used to classify generative AI systems to conform to risk classifications used by authorities. This is the goal of our current research project at Howest, Belgium, on “AI in design” (AID) for 3 application domains (creative agencies, industrial product design and marketing/communication). Creating a “moral compass” regarding the use of generative AI systems in specific contexts, where developers and users are able to evaluate the moral standing of AI products and avoid the creation of a censoring system but stimulating the human ethical reflection.

Speakers

Geert Hofman

Lecturer/Researcher, Howest, University College of West-Flanders

After a career in industry as a developer, I started my own company in 1995 building a local ISP. In 2005 I sold my company. Since then I am a lecturer at Howest in IT and critical thinking related topics. Research is also part of my job description. Topics range from open source... Read More →

Wednesday June 19, 2024 12:05 - 12:35 CEST
Saint Germain (Level 3)

Ethical AI Practices

Audience Experience Level Beginner

12:35 CEST

Lunch Break (Provided Onsite for All Attendees)

Wednesday June 19, 2024 12:35 - 13:50 CEST
Solutions Showcase, Theatre (Level -1)

Special Events / Exhibits / Breaks

12:35 CEST

Women and Non-Binary Lunch

We’d like to invite all attendees who identify as women or non-binary to join each other for a networking lunch at the event. We will begin with a brief introduction and then attendees will be free to enjoy lunch and mingle with one another. All attendees must identify as a woman or non-binary and must be registered for the conference to attend.
*We will do our best to accommodate all interested attendees, but please note that participation is on a first-come, first-served basis.

Wednesday June 19, 2024 12:35 - 13:50 CEST
TBA

Special Events / Exhibits / Breaks

13:50 CEST

AI and Incident Management: How to Reduce Manual Work and Speed up the Incident Response Process - Birol Yildiz, ilert GmbH

This talk explores the transformative impact of Artificial Intelligence on IT incident management, emphasizing its potential to drastically reduce manual workload and accelerate response times during critical incidents.

Speakers

Birol Yildiz

CEO, ilert GmbH

Birol Yildiz is the Co-founder and CEO of ilert, adeptly steering the company with a rare combination of technical and product expertise. His prior experience includes a significant role as Chief Product Owner for Big Data products at REWE Digital. With a strong foundation in Computer... Read More →

Wednesday June 19, 2024 13:50 - 14:20 CEST
Jussieu (Level 3)

AI in Action

Audience Experience Level Any

13:50 CEST

Dive into the Lightning AI Open Source Stack and Lightning Studios to Unlock Reproducible AI Development on the Cloud - Luca Antiga, Lightning AI

Dive into the Lightning AI open source stack and Lightning Studios to unlock reproducible AI development on the cloud.
PyTorch Lightning is a leading open source framework that was used to train several of the best generative AI models. With over 100 million downloads, it is the framework of choice for researchers and companies worldwide to train and fine-tune AI models. PyTorch Lightning and the rest of the Lightning AI stack, which includes Fabric, TorchMetrics, litgpt, litdata, and Thunder, provides a cutting-edge open-source foundation for practitioners.
Just as the Lightning AI open source stack is democratizing access to cutting-edge AI research and engineering, Lightning Studios democratize access to cloud computing resources and solve the challenges of reproducibility and collaboration. Lightning Studios offer a laptop-like cloud experience, enabling seamless accelerated computing for individuals and organizations. Studios provide reproducible environments for development, training, and hosting applications across diverse hardware setups, with support for multi-node training and parallel data processing.
Attendees will gain firsthand experience and access to datasets, models, and source code.

Speakers

Luca Antiga

CTO, Lightning AI

Luca Antiga is the CTO at Lightning AI. He is an early contributor to PyTorch core and co-authored “Deep Learning with PyTorch” (published by Manning). He started his journey as a researcher in Bioengineering and later co-founded Orobix, a company focused on building and deploying... Read More →

Wednesday June 19, 2024 13:50 - 14:20 CEST
Sorbonne Descartes & Lutèce (Level 5)

AI Research & Methodologies

13:50 CEST

Open-Sourcing Highly Capable Foundation Models: Evaluation of Risks, Benefits, and Alternative Means - Abhishek Gupta, BCG Henderson Institute

Recent decisions by leading AI labs to either open-source their models or to restrict access have sparked debate about whether and how increasingly capable AI models should be shared. Open-sourcing in AI typically means making model architecture and weights freely and publicly accessible for anyone to modify, study, build on, and use. This offers advantages such as enabling external oversight, accelerating progress, and decentralizing control over AI development and use. However, it also presents a growing potential for misuse and unintended consequences. This paper examines the risks and benefits of open-sourcing highly capable foundation models. While open-sourcing has historically provided substantial net benefits for most software and AI development processes, I argue that for some highly capable foundation models likely to be developed in the near future; open-sourcing may pose sufficiently extreme risks to outweigh the benefits. In such a case, highly capable foundation models should not be open-sourced, at least not initially. Alternative strategies, including non-open-source model-sharing options, are explored.

Speakers

Abhishek Gupta

Fellow, Augmented Collective Intelligence, BCG Henderson Institute

Abhishek Gupta is the Director for Responsible AI with BCG and Founder & Principal Researcher at the Montreal AI Ethics Institute. His work focuses on applied technical, policy, and organizational measures in Responsible AI. Before, he worked as an ML Engineer at Microsoft and served... Read More →

Wednesday June 19, 2024 13:50 - 14:20 CEST
Saint Germain (Level 3)

Ethical AI Practices

Audience Experience Level Intermediate

13:50 CEST

Workshop: Deploy and Monitor ML Pipelines with Open Source and Free Applications - Rami Krispin, Apple

The workshop will focus on different deployment designs of machine learning pipelines using open-source applications and free-tier tools. We will use the US hourly demand for electricity data from the EIA API to demonstrate the deployment of a pipeline with GitHub Actions and Docker that fully automates the data refresh process and generates a forecast on a regular basis. This includes the use of open-source tools such as MLflow and YData Profiling to monitor the health of the data and the model's success. Last but not least, we will use Quarto doc to set up the monitoring dashboard and deploy it on GitHub Pages.

Speakers

Rami Krispin

Senior Manager Data Science and Engineering, Independent

Rami Krispin is a data science and engineering manager who mainly focuses on time series analysis, forecasting, and MLOps applications. He is the author of Hands-On Time Series Analysis with R and is currently working on my next book, Applied Time Series Analysis and Forecasting... Read More →

Wednesday June 19, 2024 13:50 - 15:05 CEST
Saint-Victor (Level 3)

AI Quality & Security

Audience Experience Level Intermediate

13:50 CEST

Workshop: Working with Gemma and Open LLMs on Google Kubernetes Engine - Abdel Sghiouar & Victor Dantas, Google Cloud

The Gemma family of open models can be fine-tuned on your own custom dataset to perform a variety of tasks, such as text generation, translation, and summarization. Combined with Kubernetes, you can unlock the open source AI innovations with scalability, reliability, and ease of management. In this workshop, you will learn through a guided hands-on exercise how you can work with Gemma and fine-tune it on a Kubernetes cluster. We will also explore options for serving Gemma on Kubernetes with accelerators and Open Source tools.

Speakers

Victor Dantas

Generative AI Solutions Architect, Google Cloud

Victor joined Google Cloud in 2022 as a Customer Engineer, having previously worked as a Solutions Architect at Amazon Web Services (AWS) and as a Cloud Architect in consulting companies. In 2024, Victor transitioned to a Generative AI Solutions Architect role, where he drives generative... Read More →

Abdel Sghiouar

Cloud Native Developer Advocate, Google Cloud

Abdel Sghiouar is a senior Cloud Developer Advocate @Google Cloud. A co-host of the Kubernetes Podcast by Google and a CNCF Ambassador. His focused areas are GKE/Kubernetes, Service Mesh and Serverless.

Wednesday June 19, 2024 13:50 - 15:05 CEST
Monge (Level 3)

AI Systems & Performance

Audience Experience Level Beginner

14:35 CEST

Code Security Reinvented: Navigating the Era of AI - Joseph Katsioloudes, GitHub

Artificial intelligence (AI) already serves as a copilot in our daily lives, acting as a digital assistant and delivering personalized experiences. Despite progress in many areas, AI has historically fallen short of improving software development practices. This changed with the introduction of AI pair programmers, which distill the collective technical know-how of the world’s developers, and their widespread adoption has been quite telling. While the process of building software has become easier and faster, the question remains: What about more secure? In this session, we’ll demonstrate several ways developers can use AI to leverage the world's security knowledge through dozens of practical demos in GitHub Copilot. The audience will gain a deep understanding of AI capabilities, along with insights and best practices drawn from the lessons we learned as developers striving to ship secure code.

Speakers

Joseph Katsioloudes

Developer Advocate, GitHub

Joseph is a security expert who empowers developers to ship secure software through his research and education work at the GitHub Security Lab. His recent contributions include video content with combined 1M+ views packed with practical security tips, and the free game gh.io/securecodegame... Read More →

Wednesday June 19, 2024 14:35 - 15:05 CEST
Jussieu (Level 3)

AI in Action

Audience Experience Level Beginner

14:35 CEST

Wikidata Knowledge Graph to Enable Equitable and Validated Generative AI - Jonathan Fraine & Lydia Pintscher, Wikimedia Deutschland

Generative AI has changed the information ecosystem. Open-access knowledge graphs like Wikidata (CC0) enable open-source applications to provide authenticated information as an equitable representation of world knowledge to LLM and ML applications. The Wikidata knowledge graph represents over 100M crowd-sourced, structured data items, populated by the Wikidata communities in ~300 languages. We will present a practical integration of Wikidata's open-source, open-access knowledge graph to provide well-cited information for Generative AI inference and validation. We will dive into the capabilities for semantic search over the Wikidata knowledge graph to augment and validate generative AI inference. We will exhibit how to leverage a vectorised Wikidata search API to enhance generative AI applications with crowd-sourced data, which represents a more equitable spectrum of human knowledge than standard internet corpora.

Speakers

Lydia Pintscher

Portfolio Lead for Wikidata, Wikimedia Deutschland

Lydia Pintscher is the Portfolio Lead for Wikidata at Wikimedia Deutschland and supports the project since its inception in 2012. She studied computer science at KIT and is a long-time contributor to open source and open culture projects, including as the Vice-President of KDE e... Read More →

Jonathan Fraine

Director of Engineering, Co-Lead of Software Development, Wikimedia Deutschland

Jonathan Fraine is the Co-Lead of Wikimedia Deutschland's (WMDE) Software Development Department (SWE). He coordinates the ops and org strategy for open-source Web and AI development at WMDE. He attained a PhD in astrophysics in 2015 and built Generative AI solutions for NASA to calibrate... Read More →

Wednesday June 19, 2024 14:35 - 15:05 CEST
Sorbonne Descartes & Lutèce (Level 5)

AI Research & Methodologies

Audience Experience Level Intermediate

14:35 CEST

Panel Discussion: Navigating the Ethical Landscape: Responsible AI in Practice - Anni Lai, Futurewei; Pedro Ortiz Suarez, Common Crawl Foundation; Mirko Boehm, Linux Foundation Europe; Oita Coleman, Servant Voice Technologies; Adrian Gonzalez-Sanchez, Uno

As artificial intelligence continues to permeate various aspects of our lives, the need for responsible AI practices becomes increasingly evident. This panel discussion aims to delve into the multifaceted dimensions of responsible AI, exploring the ethical considerations, societal impacts, and practical approaches in deploying AI technologies. Join our diverse panel of experts as they share insights, discuss challenges, and propose solutions to foster a more responsible and ethically sound AI ecosystem.

Speakers

Adrián González Sánchez

Data & AI Specialist, AI Ethics Lead, Linux Foundation / Microsoft / OdiseIA

Data & AI Specialist at Microsoft, Member at LF Generative AI Commons (RAI Workstream) and OdiseIA (Spanish AI Observatory).University Professor at HEC Montreal and IE University, 2 x Book Author at O'Reilly (KCNA Study Guide & Azure OpenAI), Online instructor at Linux Foundation... Read More →

Richard Sikang Bian

Senior Manager, New Strategic Initiatives / Head of Open Source Program Office, Ant Group

Richard Sikang Bian is currently working at Ant Group’s Technical Strategic Development team. As an engineer by training, Richard was an ex-Square, ex-Microsoft seasoned software engineer who had been living in the States for 10+ years before joining Ant Group.Richard’s current... Read More →

Anni Lai

Head of Open Source Operations. Chair of Generative AI Commons, Futurewei

Anni drives Futurewei’s open source (O.S.) governance, process, compliance, training, project alignment, and ecosystem building. Anni has a long history of serving on various O.S. boards such as OpenStack Foundation, LF CNCF, LF OCI, LF Edge, and is on the LF OMF board and LF Europe... Read More →

Mirko Boehm

Sr Director of Community Development, Linux Foundation Europe

Community Development at Linux Foundation Europe

Oita Coleman

CEO/Founder, Servant Voice Technologies

Pedro Ortiz Suarez

Senior Research Scientist, Common Crawl Foundation

Pedro is a senior research scientist at the Common Crawl Foundation. He holds a PhD in computer science and Natural Language Processing from Sorbonne Université. Pedro’s research has mainly focused on how data quality impacts ML models’ performance and how to improve these models... Read More →

Wednesday June 19, 2024 14:35 - 15:05 CEST
Saint Germain (Level 3)

Ethical AI Practices

Audience Experience Level Beginner

15:05 CEST

Coffee Break

Wednesday June 19, 2024 15:05 - 15:35 CEST
Grande Galerie (Level -1)

Special Events / Exhibits / Breaks

15:35 CEST

Between Public Benchmarks and Business Needs - Lukasz Borchmann, Snowflake

How do you choose the best LLM for your task? We will seek practical insights into LLM applications by examining the tradeoffs in LLM design, questioning widespread evaluation standards, and learning from Snowflake Arctic's development.

Speakers

Lukasz Borchmann

Senior Research Scientist, Snowflake

Machine Learning researcher specializing in Natural Language Processing and Document Understanding. With a strong background in the industry and several international competitions won, he has contributed to the advancement of language modeling, particularly in multi-modal models incorporating... Read More →

Wednesday June 19, 2024 15:35 - 16:05 CEST
Jussieu (Level 3)

AI in Action

15:35 CEST

Designing AI for Network Troubleshooting: A Langchain Proof of Concept - Jesus Illescas, Cisco

Network downtime is a costly issue that demands efficient troubleshooting. This session introduces a proof of concept using AI and Langchain for AI-enhanced network troubleshooting. By interfacing the Language Model (LLM) with network devices, we extract crucial network state information for effective troubleshooting. Attendees will gain insights into the process of designing AI systems for network troubleshooting and witness a live demonstration of the concept. This session is ideal for developers and network professionals with an intermediate understanding of Python.

Speakers

Jesus Illescas

Developer Advocate, Cisco

Tech enthusiast, passionate about SW engineering, DevOps & Networks. Currently working as Developer Advocate with Cisco DevNet.

Wednesday June 19, 2024 15:35 - 16:05 CEST
Sorbonne Descartes & Lutèce (Level 5)

AI in Action

Audience Experience Level Intermediate

15:35 CEST

Examining the Principles of Observability and Its Relevance in LLM Applications - Guangya Liu & Nimesh Bhatia, IBM

Large Language Models (LLMs) represent a significant leap in artificial intelligence, trained on extensive text and code datasets to perform tasks like text generation, language translation, and responsive questioning. While still evolving, LLMs are already being used in a variety of applications, including chatbots, search engines, and creative writing tools, thus monitoring and understanding AI behaviors becomes important. Users demand transparency, not a "black box," seeking insights into the AI's decision-making processes. Observability addresses this by gathering and analyzing data to refine LLM performance, uncover biases, troubleshoot problems, and guarantee AI reliability and trustworthiness. In this session, we will deep dive to LLM Observability, including metrics should be observed, like model latency/cost, model tracking etc , emerging technologies including traceloop, otel, langfuse etc, how to use those technologies to do analytics, monitoring and optimization for LLM apps.

Speakers

Nimesh Bhatia

Director, IBM Instana Observability, IBM

Nimesh is an engineering leader at IBM's Instana Observability. In the past he has led open source engineering org which contributes to many strategic CNCF projects such as Kubernetes, Knative, Istio and many more. He has over 10 patents and many IBM internal and external publica... Read More →

Guangya Liu

Senior Technical Staff Member, IBM

Guangya Liu is a Senior Technical Staff Member (STSM) for IBM Instana. He is the technical leader driving the IBM Instana foundation development and customer support. He is also the open source leader for the overall IBM Instana team, driving the cloud-native open source contribution... Read More →

Wednesday June 19, 2024 15:35 - 16:05 CEST
Saint-Victor (Level 3)

AI Quality & Security

Audience Experience Level Beginner

15:35 CEST

Introduction to Distributed ML Workloads with Ray on Kubernetes - Abdel Sghiouar, Google Cloud

The rapidly evolving landscape of Machine Learning and Large Language Models demands efficient scalable ways to run distributed workloads to train, fine-tune and serve models. Ray is an Open Source framework that simplifies distributed machine learning, and Kubernetes streamlines deployment. In this introductory talk, we'll uncover how to combine Ray and Kubernetes for your ML projects. You will learn about: - Basic Ray concepts (actors, tasks) and their relevance to ML - Setting up a simple Ray cluster within Kubernetes - Running your first distributed ML training job

Speakers

Abdel Sghiouar

Cloud Native Developer Advocate, Google Cloud

Wednesday June 19, 2024 15:35 - 16:05 CEST
Monge (Level 3)

AI Systems & Performance

Audience Experience Level Beginner

15:35 CEST

From OpenAI to Opensource AI: Navigating Between Commercial Ownership and Collaborative Openness - Raphaël Semeteys, Worldline

This presentation explores the evolution of generative AI, highlighting the trajectories of various models such as GPT-4, and examining the dynamics between commercial interests and the ethics of open collaboration. We offer an in-depth analysis of the levels of openness of different language models, assessing various components and aspects, and exploring how the (de)centralization of computing power and technology could shape the future of AI research and development. Additionally, we explore concrete examples like LLaMA and its descendants, as well as other open and collaborative projects, which illustrate the diversity and creativity in the field, while navigating the complex waters of intellectual property and licensing.

Speakers

Raphaël Semeteys

Head of DevRel, Architect, Open Source Expert, Worldline

25 years experience in IT in several business fields and positions (dev, run, business analyst, project manager, architect, consulting, presales... and now DevRel)Strong expertise about free and open source software (9 years in a dedicated skill center, animating communities, talks... Read More →

Wednesday June 19, 2024 15:35 - 16:05 CEST
Saint Germain (Level 3)

Ethical AI Practices

Audience Experience Level Beginner

16:20 CEST

Rethinking Software and Software Engineering in the Era of Foundation Models (SE4FMware) - Ahmed E. Hassan, SAIL - Queen's University

Foundation models (FMs), such as Large Language Models (LLMs), have revolutionized software development by enabling new use cases and business models. We refer to software built using FMs as FMware. The unique properties of FMware (e.g., prompts, agents, and the need for orchestration), coupled with the intrinsic limitations of FMs (e.g., hallucination) lead to a completely new set of software engineering challenges. Based on our extensive experience in this domain in particular and Software Engineering in general, we identified 10 key SE4FMware challenges that have caused FMware development to be unproductive, costly, and risky. In this talk, we discuss some of these challenges in detail and state the path for innovation that we envision. We hope that the disclosure of the aforementioned challenges and our associated efforts to tackle them will not only raise awareness but also promote deeper and further discussions, knowledge sharing, and innovative solutions across the software engineering discipline. More info at: https://fmse.io/ and https://2024.aiwareconf.org and ttps://arxiv.org/abs/2402.15943

Speakers

Ahmed E. Hassan

Professor, SAIL - Queen's University

Hassan is an ACM/IEEE/NSERC Steacie fellow. He is a laureate of the Mustafa Prize, a distinction equivalent in prestige to a Nobel, for founding the AI-augmented SE field and its associated Mining Software Repositories (MSR) conference. He is a Distinguished Educator of IEEE TCSE... Read More →

Wednesday June 19, 2024 16:20 - 16:50 CEST
Jussieu (Level 3)

AI in Action

Audience Experience Level Any

16:20 CEST

The Dark Side of AI: The Hidden Supply Chain Risks in Open-Source AI Models - Jossef Harush Kadouri & Tzachi Zornshtain, Checkmarx

There's a rise in powerful AI tools. Are they secure enough? In this talk, I will demonstrate how open source AI models are built, with a focus on their weak spots. I will then give a live demo of how to exploit these weaknesses to build malicious models and show how easy it is to publish them to HuggingFace. I will show the following techniques: - embedding malicious code in models - data poisoning for existing models and training to target specific victims

Speakers

Tzachi Zornshtain

Head of SCS, Checkmarx

Zack heads the Supply Chain Security product unit at Checkmarx. With vast malware research experience, he brings invaluable knowledge and skills to the table. Prior to Checkmarx, Zack co-founded Dustico in 2020, a software supply chain security company acquired by Checkmarx in 2021... Read More →

Jossef Harush Kadouri

Head of Supply Chain Security, Checkmarx

In 2020, I co-founded Dustico (acquired by Checkmarx), a software supply chain security company. Since then, I have been working with my team to identify and prevent software supply chain attackers. I have a passion for startups: Dustico (as a co-founder), Zero Networks (as a co-founder... Read More →

Wednesday June 19, 2024 16:20 - 16:50 CEST
Saint-Victor (Level 3)

AI Quality & Security

Audience Experience Level Beginner

16:20 CEST

AI Software Bill of Materials: From the EU AI Act to Your MLOps - Adrian Gonzalez Sanchez, Linux Foundation / Microsoft / OdiseIA

The EU AI Act brings new requirements for AI systems, including transparency measures for adopters and providers. Regardless of the applicability or impact of this AI regulation to the open source ecosystem, the introduction of a new AI Software Bill of Materials is a great opportunity to structure the way companies are transparent, by leveraging a json file that integrates with their MLOps processes.

Speakers

Adrián González Sánchez

Data & AI Specialist, AI Ethics Lead, Linux Foundation / Microsoft / OdiseIA

Wednesday June 19, 2024 16:20 - 16:50 CEST
Saint Germain (Level 3)

AI Regulation & Standards

Audience Experience Level Intermediate

17:05 CEST

Moxin: The All-In-One Tool to Run LLMs on Your Personal Device - Michael Yuan, Second State

Moxin is a new and fast growing open-source project to run Large Language Models (LLMs) on your personal computers. It has a Rust-based GUI for rich interactions from LLM model exploration, selection, to chatting. Using the Rust in-process messaging channels, it has a backend service that runs LLM apps, and respond to user messages. The backend service is based on the WasmEdge Runtime, which enables the application to take advantage of a variety of the different GPUs or other accelerators that might be available on the device. In addition, the team has also built an LLM database that provides up-to-date meta-data and benchmark results services to all open-source LLM models on Hugging Face.

In this talk, we will showcase the Moxin app and discuss the design decisions behind it. The application offers a template and a component library for Rust developers to create their own cross-platform LLM applications. Developer will learn how to build rich UI applications, and LLM services in Rust.

Speakers

Michael Yuan

Maintainer, CNCF WasmEdge and CEO, Second State

Dr. Michael Yuan is a maintainer of WasmEdge Runtime (a project under CNCF) and a co-founder of Second State. He is the author of 5 books on software engineering published by Addison-Wesley, Prentice-Hall, and O'Reilly. Michael is a long-time open-source developer and contributor... Read More →

Wednesday June 19, 2024 17:05 - 17:35 CEST
Jussieu (Level 3)

AI in Action

Audience Experience Level Beginner

17:05 CEST

Introducing the Model Openness Framework: Achieving Completeness and Openness in a Confusing Generative AI Landscape - Anni Lai, Futurewei; Matt White & Cailean Osborne, Linux Foundation

Join us in this enlightening session as we delve into the dynamic realm of generative AI and explore the groundbreaking Model Openness Framework introduced by Generative AI Commons. In an era where artificial intelligence is rapidly evolving, the significance of openness cannot be overstated. Our talk will unravel the layers of openness in generative AI, shedding light on the ethical, societal, and technological aspects that make transparency and collaboration crucial. We will navigate through the intricacies of the Model Openness Framework, a pioneering initiative that sets the stage for a new era of responsible and inclusive AI development. Discover how this framework not only facilitates the sharing of knowledge but also fosters a community-driven approach to harness the potential of generative AI. Whether you're a seasoned AI enthusiast or a newcomer to the field, this talk promises valuable insights into the transformative power of openness in shaping the future of generative AI. Let's embark on a journey together, exploring the ethical dimensions, societal impacts, and technical considerations that define the landscape of openness in the world of AI.

Speakers

Cailean Osborne

Researcher, LF Research

Cailean is a Researcher at the Linux Foundation and a PhD Candidate in Social Data Science at the University of Oxford. His PhD thesis concerns the political economy of open source AI. Previously, Cailean worked as the International Policy Lead at the UK Government’s Centre for... Read More →

Anni Lai

Head of Open Source Operations. Chair of Generative AI Commons, Futurewei

Matt White

Executive Director, PyTorch Foundation & GM of AI, The Linux Foundation

Matt White is the Executive Director of the PyTorch Foundation and GM of AI at the Linux Foundation. He is also the Director of the Generative AI Commons, an open community initiative focused on advancing responsible generative AI under the LF AI & Data Foundation. Matt has nearly... Read More →

Wednesday June 19, 2024 17:05 - 17:35 CEST
Saint Germain (Level 3)

AI Regulation & Standards

Audience Experience Level Intermediate

17:30 CEST

Onsite Reception

When day 1 sessions conclude, all attendees are invited to gather at the Onsite Reception to connect over refreshing drinks and hors d'oeuvres. Head to the Solutions Showcase to network, explore cutting-edge sponsor products, and check out the latest technological advancements and trends. We hope to see you there!

Wednesday June 19, 2024 17:30 - 19:00 CEST
Solutions Showcase, Theatre (Level -1)

Special Events / Exhibits / Breaks

06:45 CEST

5K Fun "Run"

Pack your running shoes because the Fun “Run” is on! This activity is great for all fitness levels with two different pace groups: a walking and a running group. Runners will get to see some local Parisian sights while getting their morning workout in!
There is no cost to participate and space is available on a first-come, first-served basis.
*Participants must be registered for AI_dev Europe 2024, and will be required to provide their own running attire and water.

Thursday June 20, 2024 06:45 - 08:00 CEST
Meet outside Maison de la Mutualité

Special Events / Exhibits / Breaks

07:30 CEST

Welcome Coffee

Thursday June 20, 2024 07:30 - 09:00 CEST
Grande Galerie (Level -1)

Special Events / Exhibits / Breaks

07:30 CEST

Coat & Bag Check

Thursday June 20, 2024 07:30 - 17:30 CEST
Ground Floor Foyer

Special Events / Exhibits / Breaks

08:00 CEST

Registration & Badge Pick-up

Thursday June 20, 2024 08:00 - 17:00 CEST
Ground Floor Foyer

Special Events / Exhibits / Breaks

09:00 CEST

Keynote: Welcome Back - Ibrahim Haddad, Executive Director, LF AI & Data Foundation

Speakers

Ibrahim Haddad

Executive Director, LF AI & Data Foundation

Thursday June 20, 2024 09:00 - 09:05 CEST

Keynote Sessions

09:05 CEST

Keynote: Eiso Kant, Co-Founder & CTO, poolside

Speakers

Eiso Kant

CTO & Co-founder, poolside

Thursday June 20, 2024 09:05 - 09:20 CEST
Theatre (Level -1)

Keynote Sessions

09:25 CEST

Keynote: Anastasia Stasenko, Co-founder, pleias & Associate Senior Lecturer, Sorbonne-Nouvelle

Speakers

Anastasia Stasenko

Co-founder, pleias & Associate Senior Lecturer, Sorbonne-Nouvelle

Anastasia Stasenko is a cofounder of pleias, a French startup specialised in development of open science LLMs trained on the fully open copyright-free data. In parallel, she holds a position of Senior Associate Lecturer in Data Analysis and Digital Strategy at Sorbonne-Nouvelle U... Read More →

Thursday June 20, 2024 09:25 - 09:40 CEST
Theatre (Level -1)

Keynote Sessions

09:45 CEST

Keynote: Omar Sanseviero, Chief Llama Officer & Head of Platform and Community, Hugging Face

Speakers

Omar Sanseviero

Chief Llama Officer and Head of Platform and Community, Hugging Face

Omar Sanseviero is the Chief Llama Officer and Head of Platform and Community at Hugging Face, where he works at the intersection of open source, community, and product. Omar leads multiple ML teams that work on topics such as Mobile ML, ML for art, and ML Partnerships. Previously... Read More →

Thursday June 20, 2024 09:45 - 10:00 CEST
Theatre (Level -1)

Keynote Sessions

10:05 CEST

Keynote: Niall Turbitt, Senior Staff Data Scientist, Databricks

Speakers

Niall Turbitt

Senior Staff Data Scientist, Databricks

Niall Turbitt is a Senior Staff Data Scientist on the ML Practice team at Databricks. In this role he works with customers to architect, build, and productionize ML applications on Databricks. He specializes in LLM-based applications, deep learning and MLOps.

Thursday June 20, 2024 10:05 - 10:20 CEST
Theatre (Level -1)

Keynote Sessions

10:25 CEST

Keynote: New Advances for Cross-platform AI Applications in Docker - Michael Yuan, Maintainer, CNCF WasmEdge and CEO, Second State & Justin Cormack, Chief Technology Officer, Docker, Inc

We will discuss new ways to support cross-platform GPU / AI workloads in container ecosystems, specifically with Docker’s support for the WebGPU standard. WebGPU allows container applications to access the host GPUs and other AI accelerators through a portable API. There is no more need to build Docker images for GPU vendors and their proprietary drivers. We will demonstrate how the WasmEdge project builds on the WebGPU standard to create portable LLM inference applications in Rust, and have those apps seamless managed and orchestrated by Docker.

Community impact:

A major advantage of open-source LLMs is to run them on each user’s private computers. Those include personal laptops, AI PCs, private or hybrid or edge cloud servers, edge devices, or even mobile devices. However, the heterogeneous hardware and software accelerators on those private computers or devices also pose great challenges for today’s AI developers. For example, today, an LLM app requires separate Docker images for all combinations of Nvidia CUDA, CUDNN, AMD ROCm, or Intel AVX etc. Each of those platform-dependent Docker images need to be developed, tested, and deployed separately.

WebGPU and Wasm provide an opportunity to abstract away and unify these underlying AI hardware and software stacks. With Docker support, they allow developers to create truly portable AI applications, and allow ops to manage those applications using their familiar Docker and Kubernetes tools.

Speakers

Justin Cormack

Chief Technology Officer, Docker, Inc

Justin is the CTO at Docker, and a member of the CNCF TOC. He has spent a lot of time working on security in the container ecosystem. He is a maintainer on the Notary project for container security.

Michael Yuan

Maintainer, CNCF WasmEdge and CEO, Second State

Thursday June 20, 2024 10:25 - 10:40 CEST
Theatre (Level -1)

Keynote Sessions

10:40 CEST

Coffee Break

Thursday June 20, 2024 10:40 - 11:20 CEST
Grande Galerie (Level -1)

Special Events / Exhibits / Breaks

11:20 CEST

Hugging Face Accelerate: Making Device-Agnostic ML Training and Inference Easy at Scale - Zachary Mueller, Hugging Face

Hugging Face Accelerate is an open-source library designed to make machine learning model training and inference simple and device-agnostic through an easy-to-use interface. It does so by keeping things at a low level, minimising the abstraction while maximising the freedom a user has over their code. Over the last two years, it has garnished new features to enhance usability in the ecosystem, such as; training agnostically between all of the latest ML acceleration hardware (CUDA, XLA, NPU, and XPU), training support in lower precisions for speed and memory efficiency, and easy-to-deploy large model inference that can scale with the amount of compute available through features like big model inference and pipeline parallelism. In this presentation, we will first cover some background into just what Accelerate is as a framework. Building on this overview, we will then detail how it has impacted the ML world at large. By the end of this talk, we will not only have identified the core parts of Accelerate but also introduce you to its easy-to-use device-agnostic API to help kickstart the journey into large-scale computing and local-first deployment of machine learning models.

Speakers

Zachary Mueller

Hugging Face

Zach Mueller is the Technical Lead for the Accelerate project at Hugging Face. He's a graduate of the University of West Florida and has considerable experience with Hugging Face as well as the fastai communities.

Thursday June 20, 2024 11:20 - 11:50 CEST
Jussieu (Level 3)

AI in Action

Audience Experience Level Any

11:20 CEST

Protecting Open Innovation in Future Regulation - Ben Brooks, Stability AI

Open models play a vital role in promoting transparency, competition, and security in AI. However, in their race to introduce new rules for model development, governments run the risk of stifling open innovation. Join Ben Brooks, Head of Public Policy at Stability AI, to explore how recent developments in AI regulation may affect the open developer ecosystem across the EU, US, and beyond. Each jurisdiction is taking a radically different approach to reform with direct and indirect consequences for those who train, tune, and deploy open models. Hot on the heels of the EU's AI Act coming into force, this session will compare how different governments think about the opportunities and risks of open innovation in AI; what different regulatory frameworks mean for open models; and how developers and researchers can get involved to help to shape AI reform.

Speakers

Ben Brooks

Head of Public Policy, Stability AI

Ben Brooks is Head of Public Policy for Stability AI. Engaging policymakers around the world, Ben works to promote transparency and competition in AI through open innovation in models. Ben has testified before the US Congress and UK Parliament, engaged closely with the EU to protect... Read More →

Thursday June 20, 2024 11:20 - 11:50 CEST
Sorbonne Descartes & Lutèce (Level 5)

AI Regulation & Standards

11:20 CEST

A Hopeful View of Truly Open Source AI - Julia Ferraioli & Tom Callaway, AWS

When it comes to artificial intelligence, people tend to fall into one of two camps: that of unbridled excitement or that of doom and gloom. It is a divisive topic, made all the more so by the many layers of indirection involved in building machine learning technologies. On the one hand, we all interact with machine learning every day in generally positive ways. On the other hand, we have also seen how easy it is for intelligent agents to Go Wrong™. But what if we had full transparency into machine learning? What if we had the same capabilities with machine learning systems as we have come to expect with open source software– the ability to study them, ascertain provenance, understand (and participate in!) the design process, and raise visibility of deficits? The effect of such systems could go far beyond democratizing machine learning and creating better intelligent systems. Let's take an optimistic look at some of the possible unexpected societal benefits that could result from truly open source AI.

Speakers

Tom Callaway

Open Source Strategist, Hacker of Things, AWS

codename: spot affiliations: AWS, Red Hat (former), Fedora publications: Raspberry Pi Hacks (2013, O’Reilly), Fedora Packaging Guidelines, Fedora Legal Guidelines affinities: 3d printing, pinball, hockey, games, geocaching, fine drinks, b-movies, scifi, trivia, traveling, frogs... Read More →

Julia Ferraioli

Open Source AI/ML Strategist, AWS

julia traces her love of open source back to her time working in machine learning research. Since then, she has contributed to the release, care, and community of open source projects and the open source ecosystem at large. From helping develop strategy for projects to securing infrastructure... Read More →

Thursday June 20, 2024 11:20 - 11:50 CEST
Saint Germain (Level 3)

Ethical AI Practices

Audience Experience Level Beginner

11:20 CEST

Workshop: Choosing the Best Open Source LLM for Your Application - Nikolai Liubimov, HumanSignal

The pace of innovation for open source LLMs is exciting—new models are being released and fine-tuned on a daily basis. But when it comes to selecting the right LLM to solve real-world problems, the hype and open leaderboards can be misleading. Before you make significant investments in LLM infrastructure, building new GenAI pipelines, and fine-tuning, we'll walk you through the process of evaluating and selecting the best LLMs for your use case. Topics covered in the technical workshop will include: - How the open leaderboards work today: popular benchmarks and methodologies for NLP model evaluation - Dimensions to evaluate LLMs, and which measures are most important based on your use case - Why auto evaluators are not enough, and ultimately human supervision based on ground truth data is the best indicator of quality - How to efficiently apply human supervision to LLM evaluation: open source toolchain and process - Different approaches to curating test data for deterministic vs generative AI Attendees will walk away with actionable steps, referenced reports, and open source tools to evaluate LLMs for their business applications.

Speakers

Nikolai Liubimov

CTO & Co-Founder, HumanSignal

Nikolai is the co-founder and CTO of HumanSignal, and creator of Label Studio, the most popular OSS data labeling platform with 300K+ users globally. Based on his experience deploying ML models at scale for Yandex and Huawei, he believes data quality is not only essential to success... Read More →

Thursday June 20, 2024 11:20 - 12:35 CEST
Saint-Victor (Level 3)

AI Quality & Security

Audience Experience Level Intermediate

11:20 CEST

Workshop: Efficient and Portable AI / LLM Inference on the Edge Cloud - Xiaowei Hu, Second State

As AI applications gain popularity, we are increasingly seeing requirements to run AI or even LLM workloads on the edge cloud with heterogeneous hardware (eg GPU accelerators). However, the simplistic approaches are too heavyweight, too slow and not portable. For example, the PyTorch container image is 3GB and a container image for a C++ native toolchain is 300MB. Python apps also require complex dependency packages and could be very slow. Those container images are dependent on the underlying host’s CPU and GPU, making them difficult to manage. Wasm has emerged as a lightweight runtime for cloud native applications. For an AI app, the entire Wasm runtime and app can be under 20MB. The Wasm binary app runs at native speed, integrates with k8s and is portable across CPUs & GPUs. In this tutorial, we will demonstrate how to create and run Wasm-based AI applications on edge server or local host. We will showcase AI models and libraries for media processing (Mediapipe), vision (YOLO, amd Llava) and language (Llama2 series of models). You will be able to run all examples on your own laptop at the session.

Speakers

Vivian Hu

Product Manager, Second State

Vivian Hu is a Product Manager at Second State and a columnist at InfoQ. She is a founding member of the WasmEdge project. She organizes Rust and WebAssembly community events in Asia.

Thursday June 20, 2024 11:20 - 12:35 CEST
Monge (Level 3)

AI Systems & Performance

Audience Experience Level Beginner

12:05 CEST

Empowering AI Education with Rag: Building Domain-Specific LLM Agents for Enhanced Learning - Miley Fu, WasmEdge

In AI literacy and education, the challenge lies in bridging the gap between general-purpose Large Language Models (LLMs) and the specific needs of diverse learning communities. This talk, explores the innovative application of Rust in developing Retrieval-Augmented Generation (RAG)-based ChatGPT agents. By leveraging RAG and vector databases, educators and developers can create LLM agents tailored to specific domains, such as certain programming languages like python, machine learning, or cutting edge LLM related thesis like papers on Sora, enhancing people's AI knowledge relevance and accuracy. Miley will share insights from creating a Rust-powered RAG bot, demonstrating how this approach simplifies the integration of domain knowledge into LLMs, making AI education more accessible and effective. Attendees will learn practical steps to implement their domain-specific LLM agents, fostering a more literate and empowered AI community.

Speakers

Miley Fu

Founding Member and DevRel, WasmEdge

Miley is a Developer Advocate dedicated to build and contribute to open source tech and community. With over 5 years of experience working on WasmEdge runtime in CNCF sandbox as the founding member, she talked at KubeCon, KCD Shenzhen, CloudDay Italy, DevRelCon, Open Source Summit... Read More →

Thursday June 20, 2024 12:05 - 12:35 CEST
Jussieu (Level 3)

AI in Action

Audience Experience Level Beginner

12:05 CEST

The Open Source AI Dilemma: Crafting a Clear Definition for Open Source AI - Ofer Hermoni, Linux Foundation AI & Data

Join us for a vital discussion on defining Open-Source AI, a task emerging as both essential and complex in the evolving AI landscape. In this session, we'll address the nuances and challenges unique to AI, distinguishing it from traditional Open-Source Software. Our focus will include the delicate balance between innovation and the four fundamental freedoms of Open Source: study, use, modify, and share. We'll explore the practicality of open AI frameworks and models, emphasizing the need for a pragmatic approach that allows for innovation without demanding complete reproducibility. This session invites the community to contribute to shaping a definition that aligns with the evolving nature of AI, balancing openness with the realities of technological advancement. Your participation is key in navigating this intricate journey and shaping the future of Open-Source AI.

Speakers

Ofer Hermoni

Generative AI Commons Education & Outreach Chair, Linux Foundation AI & Data

Ofer Hermoni is a distinguished entrepreneur, having launched groundbreaking startups in the patents and privacy arenas. Renowned in the artificial intelligence sphere, he was instrumental as a co-founder of the LF AI & Data. Today, he's at the helm of the Generative AI Commons' Education... Read More →

Thursday June 20, 2024 12:05 - 12:35 CEST
Sorbonne Descartes & Lutèce (Level 5)

AI Regulation & Standards

Audience Experience Level Beginner

12:05 CEST

Toward AI Democratization with Digital Public Goods - Lea Gimpel, Digital Public Goods Alliance (DPGA) & Daniel Brumund, GIZ

This session will give insights into open-source AI systems' development, use, impact and challenges and their role as digital public goods (DPG) in global majority countries. Specifically, we will discuss the different facets of AI democratization and the contribution of open-source AI toward its ends. We will also present case studies of public sector open-source AI projects from Rwanda and Kenya and discuss how open-source AI projects can flourish in under-resourced contexts.

Speakers

Daniel Brumund

Advisor - FAIR Forward „Artificial Intelligence for All“, GIZ

In his professional life, Daniel Brumund (he/him) works as technical advisor on AI and digital public goods with GIZ's initiative "FAIR Forward - Artificial Intelligence for All". Recently, he has been focusing on community-driven data collection, NLP in low-resource languages, open-source... Read More →

Lea Gimpel

Director of AI and Country Engagement, Digital Public Goods Alliance (DPGA)

Lea is the Director of AI at the Digital Public Goods Alliance. Previously, she co-led the GIZ program "FAIR Forward - Artificial Intelligence for All" which aims to democratize AI development by enabling access to open AI training datasets, amongst others. She holds two Master's... Read More →

Thursday June 20, 2024 12:05 - 12:35 CEST
Saint Germain (Level 3)

Ethical AI Practices

Audience Experience Level Any

12:35 CEST

Lunch Break (Provided Onsite for All Attendees)

Thursday June 20, 2024 12:35 - 14:15 CEST
Solutions Showcase, Theatre (Level -1)

Special Events / Exhibits / Breaks

14:15 CEST

Lightning Talk: Open Source GenAI: A Community-Powered Future - Ofer Hermoni, Linux Foundation AI & Data

In this dynamic lightning talk, we'll explore the transformative role of open-source projects in shaping the future of Generative AI (GenAI). Our journey at the Linux Foundation AI & Data has been marked by an unwavering commitment to providing the developer community with educational resources previously thought exclusive to commercial products. We'll discuss our recent efforts in understanding developers' perspectives on open-source GenAI. Highlighting the significant impact of Collaborative Innovation in GenAI, we draw parallels with the development of the Internet, a product of synergy between developers and researchers. This talk is complementary to the detailed session on 'Building Ethical AI: The Power of Open Source and Education,' where we delve deeper into the importance of open source in developing responsible and ethical AI. As we conclude, we will issue a compelling call to action, inviting attendees to join our efforts in democratizing GenAI technology.

Speakers

Ofer Hermoni

Generative AI Commons Education & Outreach Chair, Linux Foundation AI & Data

Thursday June 20, 2024 14:15 - 14:25 CEST
Saint Germain (Level 3)

Ethical AI Practices

Audience Experience Level Beginner

14:15 CEST

Advanced RAG Techniques with LlamaIndex - Pierre-Loic Doulcet, LlamaIndex

In this session we will go through the different common pain points of a RAG pipeline and how to address them.

Speakers

Pierre-Loic Doulcet

Founding AI Engineer, LlamaIndex

Founding AI engineer at LlamaIndex, Stanford Codex Affiliate.

Thursday June 20, 2024 14:15 - 14:45 CEST
Jussieu (Level 3)

AI in Action

14:15 CEST

Future-Proofing Agent Supervision - Alexandre Variengien & Diego Dorn, EffiSciences

As autonomous agents powered by LLM are deployed in the real world, there is a need for real-time monitoring to detect and mitigate their unpredictable failures. These challenges, including indirect prompt injection and strategic deception, diverge from traditional software issues due to the agents' emergent capabilities and continuous learning. The question arises: how do we ensure our monitoring systems can preemptively address unforeseen failures? This presentation advocates for rigorous evaluations of agent monitoring systems, highlighting the importance of diverse anomaly detection, engaging with more than just chat interfaces, and tackling nuanced issues like ethical boundaries. We propose a community-driven approach to refine LLM agent supervision, featuring a shared database of failure cases and a unified trace format across applications to foster collaborative innovation. Our framework introduce two metrics: i) accuracy on held-out anomaly, simulating the unforeseen failure modes that will emerge on the future, ii) its proficiency in spotting early warning signs before an harmful action. Join us in shaping the future of agent supervision, to anticipate the unexpected!

Speakers

Diego Dorn

Research Engineer, EffiSciences

Diego Dorn is a Research Engineer at EffiSciences, currently developing a supervisor for LLM agents and a benchmark to evaluate LLM monitoring systems. He draws his expertise from his many projects during his master's in Communication Systems at EPFL, from game jams and hackathons... Read More →

Alexandre Variengien

AI safety researcher, EffiSciences

Alexandre Variengien is a researcher who worked on scalable LLM interpretability at Conjecture and Redwood Research, publishing several papers in top conferences. He's now working as an independent researcher focusing his research efforts on generalist agent supervision. He is also... Read More →

Thursday June 20, 2024 14:15 - 14:45 CEST
Saint-Victor (Level 3)

AI Quality & Security

Audience Experience Level Intermediate

14:15 CEST

Cloud Native Artificial Intelligence - Top to Bottom - Ronald Petty, RX-M

Cloud Native (CN) is not new, Artificial Intelligence (AI) is not new, CNAI is new. In this session we discuss how CN is influencing and enabling the next generation of MLOps system implementations. From inference to interfaces, will expose how tools such as Kubeflow and KServe are connecting these worlds. In this session we will move between demo and field study, highlighting key integration points in the CNAI technology stack, lay out existing challenges, and look to the future as both CN and AI continue to evolve and how CNAI can match them. Attendees whom attend this session will see how they can leverage their CN knowledge to enter this space.

Speakers

Ronald Petty

RX-M

Ronald Petty is a consultant at RX-M, a global cloud native advisory and artificial intelligence training firm in the founding classes of Kubernetes Certified Service Providers (KCSP) and Kubernetes Training Providers (KTP). He has consulted, developed, and trained across many domains... Read More →

Thursday June 20, 2024 14:15 - 14:45 CEST
Monge (Level 3)

AI Systems & Performance

Audience Experience Level Intermediate

14:25 CEST

Lightning Talk: Collective Partnership Between Government, Academia, and Industry: Leading to Societal Solutions - Vini Jaiswal, TikTok

This session highlights a groundbreaking initiative where government support catalyzed the development of a research project, addressing a critical societal issue through the lens of AI and data science. Attendees will be taken on a journey through the lifecycle of a collaborative project, from its inception with governmental backing to the pivotal role of open-source platforms in facilitating research and development. The talk aims to showcase how open-source methodologies not only expedited problem-solving but also democratized access to advanced AI tools and data, enabling a wide spectrum of contributors to participate in the innovation process. The session will conclude with key takeaways and best practices from the project, leaving the audience with a blueprint for leveraging public-private-academic partnerships in AI. Session will emphasize the importance of open-source ecosystems in fostering innovation, transparency, and community engagement, to advance technology and collaborative innovations.

Speakers

Vini Jaiswal

Chair of Technical Advisory Council, Linux Foundation AI & Data

Vini Jaiswal is a renowned expert in AI and Data, acclaimed for her significant contributions to Apache Spark, MLflow, PrivacyGo and, notably, Delta Lake. Holding pivotal roles such as Chair of the Technical Advisory Committee (TAC) at Linux Foundation Data and AI, Governing Board... Read More →

Thursday June 20, 2024 14:25 - 14:35 CEST
Saint Germain (Level 3)

Ethical AI Practices

Audience Experience Level Any

14:35 CEST

Lightning Talk: The Collingridge Dilemma in AI - Angélina Gentaz, Consultant

The Collingridge dilemma highlights a significant challenge in technology policy: the difficulty of controlling a technology's impact when it is fully developed and widely adopted, versus the ease of control before its impacts are fully understood. This lightning talk explores this dillema in AI developement, emphasizing the unique challenges AI poses due to its rapid development and potential for unforeseen, widespread impacts. The goal is to foster a discussion on how policymakers and developers can collaboratively navigate the uncertain future of AI.

Speakers

Angélina Gentaz

AI Policy & Technical Governance, Consultant

Angélina has been working as a consultant for different AI safety and AI governance organisations inclunding the AI Policy Observatory of the OECD. Her focus is on emphasizing the importance of bridging the gap between technical AI safety and AI governance, using economic and scientific... Read More →

Thursday June 20, 2024 14:35 - 14:45 CEST
Saint Germain (Level 3)

Ethical AI Practices

Audience Experience Level Any

14:50 CEST

Coffee Break

Thursday June 20, 2024 14:50 - 15:15 CEST
Grande Galerie (Level -1)

Special Events / Exhibits / Breaks

15:00 CEST

Composing AI Applications as a Graph with Haystack - Tuana Celik, deepset

There are many moving parts in a fully functioning AI application, whether that be retrieval-augmented generation, agents, chatbots.. Haystack is an open source AI framework in Python, recently redesigned in an opinionated way, that takes the approach of building AI applications as a directed (cyclic) graph. In this talk, we will see how Haystack allows engineers to take full control over how they design their AI workflows, making use of pretty much all model providers and technologies out there.

Speakers

Tuana Celik

Developer Advocate, deepset

Tuana is a Developer Advocate at deepset where she focuses on the open source AI community and Haystack, deepset’s open source AI framework. She has been actively involved in helping and maintaining the Haystack developer community and Haystack’s educational materials.

Thursday June 20, 2024 15:00 - 15:30 CEST
Jussieu (Level 3)

AI in Action

15:00 CEST

Efficient and Cross-Platform LLM Inference in the Heterogenous Cloud - Michael Yuan, Second State

As AI/LLM applications gain popularity, there are increasing demands to run and scale them in the cloud. However, compared with traditional cloud workloads, AI workloads are heavily reliant on the GPU. Linux containers are not portable across different hardware devices, and traditional container management tools are not setup to re-compile applications on new devices at deployment time. Cloud native Wasm provides a new portable bytecode format that abstracts away GPUs and hardware accelerators for these applications. With emerging W3C standards like WASI-NN, you can write and test LLM applications in Rust on your Macbook, and then deploy on a Nvidia cloud server or an ARM NPU device without re-compilation or any change to the Wasm bytecode file. The Wasm apps can also be managed by existing container tools such as Docker, Podman, and K8s, making them a great alternative to Linux containers for this new workload. This talk will discuss how WasmEdge (CNCF sandbox) implements WASI-NN and supports a large array AI/LLM applications. You will learn practical skills on how to build and run LLM applications on ALL your local, edge, and cloud devices using a single binary application.

Speakers

Michael Yuan

Maintainer, CNCF WasmEdge and CEO, Second State

Thursday June 20, 2024 15:00 - 15:30 CEST
Monge (Level 3)

AI Systems & Performance

Audience Experience Level Beginner

15:45 CEST

Lightning Talk: Beyond Polygons: The Future of 3D Graphics with Implicit Models - Pierre Wilmot, Dimension Studio

Exploring the landscape of implicit representation for 3D or 4D (3D over time) scenes representation. Neural radiance field, neural signed distance fonctions, adding motion over time, relighting, and so on.

Speakers

Pierre Wilmot

Senior Software Engineer, Dimension Studio

Software enginer working at the intersection of AI and computer graphics.

Thursday June 20, 2024 15:45 - 15:55 CEST
Jussieu (Level 3)

AI in Action

Audience Experience Level Intermediate

16:05 CEST

Lightning Talk: Foundations, Frameworks and Tools For Machine Learning - Manka Velda Manjong, Independent

In this session, we're going to dive deep into the exciting world of machine learning. Picture this: we'll start by exploring the absolute core of how machine learning works, breaking down those complex algorithms into digestible bits so everyone, regardless of experience, can get a handle on them. Then, we're going to check out the heavy-hitters, like TensorFlow and PyTorch, and figure out how to use them practically – no jargon, just real-world applications. But wait, there's more! I'll be sharing a bunch of tools that have been absolute game-changers for me, from tinkering with data to deploying models. Whether you're a seasoned pro or just getting started, this session is all about making machine learning accessible and giving you some practical tools to take away and use in your own projects. Let's make the machine learning journey less intimidating and more exciting together!"

Speakers

Manka Velda Manjong

Passionate about AI, ML, and data science, I am a web developer and hold a bachelor's degree in computer engineering. My enthusiasm for creating innovative projects with technology has been a driving force in my journey. Actively engaged in communities, I've gained valuable insights... Read More →

Thursday June 20, 2024 16:05 - 16:15 CEST
Jussieu (Level 3)

AI in Action

Audience Experience Level Beginner