Opening Keynotes

Our 2025 speakers

Yann LeCun is a French computer scientist known for his pioneering work in deep learning and artificial intelligence. He developed convolutional neural networks and contributed to major advances in computer vision and pattern recognition. LeCun is a founding figure of modern AI research and has served as Chief AI Scientist at Meta as well as a professor at New York University. In 2018 he received the Turing Award together with Geoffrey Hinton and Yoshua Bengio for their contributions to deep learning.

Pim de Witte is a Dutch tech entrepreneur known for creating Medal.tv, a widely used platform for sharing gameplay clips. He started building online projects at a young age and later founded General Intuition, a company focused on advanced AI systems. He is recognized for his early success in gaming technology and his recent work in artificial intelligence.

Jérôme Monceaux is a visionary engineer and serial entrepreneur, best known for co-founding Aldebaran Robotics and leading the creation of iconic social robots Nao and Pepper. In 2021, he founded Enchanted Tools, a deep-tech startup blending robotics, AI, and 3D character design to build “enchanted” humanoid helpers.

He is a self-taught entrepreneur and has worked in the Internet and telecommunications industry since the late 1980s. In 1993 he co-founded France’s first ISP, and in 1999 he created Free – France’s first free-access ISP. 

He co-invented triple-play and the concept of the box, launching the Freebox in 2002 – a unique, state-of-the-art, multiservices box combining broadband Internet with telephony and television. Xavier has invested in telecom operators in a personal capacity in many countries outside France, including Switzerland, Ireland and Monaco. 

In 2013, he founded “42”, a not-for-profit organization that delivers free coding training based on peer-to-peer learning. This training is available in a large number of countries and the network of “42" schools currently trains over 5,000 students worldwide. In 2017, the “42” school in Paris, France, was voted the best coding school in the world. 

Xavier is also the founder of Station F – the world’s largest start-up campus – which opened its doors in 2017 and hosts a thousand start-ups in a former railway station in Paris occupying 34,000 sq.m. In november 2023, Xavier co-founded Kyutai – a non-profit, open science AI research lab based in Paris – alongside Rodolphe Saadé (Chairman and CEO of CMA CGM) and Eric Schmidt (former Chief Executive Officer and Chairman of Google). 

Resolutely committed to the democratization of AI, Kyutai is positioning itself as a leading player in AI open science. Its ambition is to share its advances with the entire AI ecosystem – the scientific community, developers, companies, society at large and decision-makers in democracies.

After 10 years of pushing audio research at Meta, Google and Kyutai, Neil joined the start-up arena with his 1s to take their models from the lab to every voice product out there.

Rémi Cadene is the Co‑Founder and Chief Executive Officer at UMA, a European robotics company building general‑purpose mobile and humanoid robots.

Before founding UMA, he worked as a research scientist at Hugging Face and has a strong background in deep learning and robotics research.

His scientific work focuses on modeling human behaviors with neural networks, neural‑architecture development, learning methods, and explainability techniques.

Aude was previously Deputy CEO of iliad Holding, serving in this capacity since 2020, and leading several major Group projects in France, Italy and Poland, covering areas ranging from marketing, IT systems and network operations through to subscriber services. 

Aude is in charge of the Group's artificial intelligence strategy, and in this role she managed the creation of the Paris-based independent research lab, Kyutai. She is also Chairman of Scaleway and Free Pro. In addition, Aude is a member of the Boards of Directors of Millicom and Monaco Télécom. Before joining the iliad Group, Aude was Chief of Staff for the CEO of Orange Wholesale & International Networks and held various roles within Orange’s B2B division. She also spent several months with Boston Consulting Group as a Visiting Associate.

Aude has an MSc in Management Science & Engineering from the University of Standford in the United States and graduated as an engineer from Ecole Polytechnique in France.

Damien Lucas joined Scaleway as CEO early 2023. Since then, he has continued to transform the iliad Group cloud provider into a major player in its sector, with a clear objective: to become the next hyperscaler. This includes the installation, at the end of 2023, of the largest AI computing power in the European cloud in DC5, iliad flagship datacenter. But also by opening up Scaleway to a dozen key sectors, from finance to retail to manufacturing, with a resolute focus on value creation. Lucas is convinced of the need for European leaders in the cloud sector, based on the key values of sovereignty, open sourceand sustainability.

A graduate of Centrale Supélec, Lucas is a seasoned entrepreneur with over 20 years; experience in the French Tech ecosystem. He began his career in the software sector, contributing to the creation of VLC Media Player, before co-founding Anevia, a company specializing in video streaming software solutions. As a driver of business development and guardian of product-first strategies, Lucas has led major projects, including the opening of international subsidiaries, consolidation of operations and an IPO. His leadership has been particularly noticed in defining product strategy and managing software development teams, and skilfully orchestrating the transition to the cloud ecosystem.

Alexandre is a co-founder at Kyutai, a non profit lab for research in artificial intelligence based in Paris. Kyutai's mission is to lead and democratize audacious and bleeding edge research in AI. We have redefined speech modeling across all applications with the speech-to-speech conversational AI Moshi, the text LLM wrapper Unmute, or the simultaneous translation model Hibiki. Before that, Alexandre was a scientist for 3 years at Facebook AI Research in Paris, where he led the development of models for audio compression and modeling (AudioCraft, MusicGen, EnCodec). Alexandre graduated in mathematics from École Normale Supérieure, and did his PhD between INRIA and Facebook AI Research Paris working on music source separation. He is most passionate about building high impact and pragmatic research in Europe.

From lab to product with European voice model

Marco Trombetti is a computer scientist, entrepreneur, and investor. In 1999, he co-founded Translated, which pioneered the symbiotic use of artificial intelligence in the language industry. Through Translated, Trombetti introduced adaptive machine translation, delivered the first commercial application of the transformer, and fostered a symbiotic relationship between professional translators and AI. Leveraging the success of Translated and other ventures, Marco co-founded Pi Campus, a venture capital firm investing in AI-applied startups, and Pi School, an educational institution dedicated to training the next generation of AI specialists.

Translation and the limits of autoregressive transformers.

Thomas Palomares is the team lead and AI Research Engineer for Centaur at Helsing. Centaur seamlessly integrates advanced AI pilots into the cockpits of existing and future fighter aircraft, transforming the precision, speed, and adaptability of modern air forces. Thomas's work focuses on leveraging scaled RL to evaluate dynamic combat situations, formulate tactics, and execute complex maneuvers in real-time, achieving human-level performance in air combat

Prior to Helsing, Thomas co-founded and served as CTO of FarmWise (2016-2024), developing autonomous agricultural robotics recognized by Forbes 30 Under 30 and TIME's 100 Best Inventions. He holds a Master's from Stanford University and studied Applied Mathematics and Computer Science at École Polytechnique, bringing deep expertise in machine learning, robotics, and real-time systems.

The first AI flying a production fighter jet

Boris Gamazaychikov is Head of AI Sustainability at Salesforce, leading global efforts to align AI innovation with climate action. Recognized by The Independent’s Climate 100 and Business Insider’s AI Trailblazers, he is a leading voice at the intersection of technology and sustainability. At Salesforce, he spearheads initiatives to reduce AI’s environmental footprint, including the AI Energy Score—the first standardized benchmark for measuring AI model energy use. With over a decade of experience tackling complex environmental challenges, Boris is also a frequent speaker and advisor. He holds a degree in Environmental Engineering from the University of Maryland and works to bridge engineering, climate science, and AI to accelerate a sustainable future.

Beyond FLOPs: Engineering transparency into AI’s carbon footprint

Dr Cédric Mahé is an epidemiologist who spent the first part of his career conducting research and  generating evidence to inform policies within WHO, The Global Fund to Fight Aids, Tuberculosis and Malaria, and UNAIDS. He joined Sanofi vaccines in 2011.

His main area of interest is to build evidence generation driven public-private partnerships with academics and public health organizations geared toward greater public health impact. He is notably the founder and President of the Foundation for Influenza Epidemiology supporting the Global Influenza Hospital Surveillance Network (www.gihsn.org), working in close collaboration with WHO. He was also the co-lead of the AIOLOS project, a multi-source data integration approach using artificial intelligence and modeling to detect and monitor respiratory virus outbreaks, co funded by the French and the German government www.aiolos-project.org.

From molecules to models: How AI is redesigning pharmaceutical R&D

Gianluca is a Research Engineer in Reinforcement Learning at InstaDeep, where he applies advanced AI to solve complex industrial challenges. His primary focus is on automated printed circuit board (PCB) routing using Deep Reinforcement Learning, and his work has also involved tackling combinatorial optimization problems, such as airport ground handling.

He completed his PhD in Reinforcement Learning at Politecnico di Milano, where his research centered on Hierarchical Reinforcement Learning and its application to mission planning for autonomous drones. His main interests lie in Reinforcement Learning and its specializations such as Hierarchical and Inverse Reinforcement Learning and the application of these techniques to real-world problems.

Solving complex combinatorial challenges with Reinforcement Learning: the PCB routing demo

Anne-Claire Baschet is the Chief Data & AI Officer at Mirakl, where she leads the company’s global AI strategy and oversees the Data & AI organization. She drives the development of Mirakl’s agentic commerce vision, implementing cutting-edge AI products—including native GenAI agents such as the Catalog Transformer—and establishing strong governance frameworks across the company. Anne-Claire previously held roles across product, data science, and AI engineering, and is recognized for bridging deep technical expertise with strategic product leadership.

Inside the Agentic Stack: How Agency Is Rewriting Enterprise

Motohiro Kashihara is a dedicated business leader and Head of Business Strategy Office for FUJITSU-MONAKA, Fujitsu’s innovative Arm-based processor.

He has a background in software development and global business strategy at a major electronics company, covering international marketing, product planning, and new business development. He later enhanced his strategic skills at a leading consulting firm, focusing on digital transformation and business innovation for enterprise clients.

With experience in both corporate technology leadership and strategic consulting, he now leads business strategy, go-to-market, partnerships, and alliances for FUJITSU-MONAKA. His approach blends technical expertise with a strong focus on customer value, turning advanced processor technology into practical solutions for next-generation computing infrastructure.

Scaling AI Sustainably with FUJITSU-MONAKA

Christophe Lacroix is Chief Innovation and Sustainability Officer at Motul. His focus now is to draw the pathway of an established and recognized brand with a historical heritage towards the energy transition and all the associated impacts both from a technical standpoint but also from a culture point of view. Christophe was previously the Chief Operations Officer at Motul.

Christophe capitalizes on an extensive experience in the chemical industry, polymer businesses and white biotechnology where he held various technical and business global positions prior to joining the lubricant industry.

Christophe holds a Ph.D in Chemical Engineering from the Ecole Polytechnique of Montreal and an Executive MBA from HEC Paris. He is also member of the Sustainable Ocean Alliance as well as an expert of the Solar Impulse Foundation. 

Beyond Air Cooling: new frontiers in AI hardware design

Pierre-Louis Theron is a seasoned internet infrastructure entrepreneur. 

Co-Founder and CEO of Alpic, building the first cloud platform for MCP servers. Alpic bring your product in the agent era. 

Pierre-Louis previously founded Streamroot, a pioneering peer-assisted Content Delivery Network for video Streaming, acquired by Lumen technology in 2019. Where he then operated the 3 largest global CDN.

A deep dive into MCP and ChatGPT apps

Micah Hill-Smith is Co-Founder & CEO of Artificial Analysis, an independent AI benchmarking and insights company that helps engineers and organizations navigate the AI landscape. Artificial Analysis evaluates model intelligence, conducts performance testing on hardware and API providers, and operates crowd-sourced arenas that have gathered millions of votes. Artificial Analysis benchmarks are regularly cited by leading AI companies including NVIDIA, Google, and OpenAI, as well as major publications such as the Wall Street Journal, the Economist, and TechCrunch. Prior to founding Artificial Analysis, Micah was a strategy consultant at McKinsey & Company.

Benchmarking the frontier: what really matters in the new AI stack

As Co-Founder of Molia, Paul Mochkovitch helps companies build secure AI agents that assist their teams. Molia develops custom AI solutions tailored to enterprise needs, ensuring both performance and security.

His background spans a decade in AI: after earning a master’s degree from Télécom Paris in 2015, he spent five years as a Data Scientist and Head of Data in tech startups before co-founding Emil, a data and AI training platform later acquired by Le Wagon. He also founded GAIN (Generative AI Network), France’s leading GenAI community with over 1,200 technical experts, and hosts Paroles d’IA, a podcast exploring enterprise AI use cases.

He is passionate about making AI accessible and actionable for businesses ready to transform their operations.

From edge inference to generative pipelines: building video AI that works

Romain Dillet is a VC investor at Drysdale Ventures based in Paris. Before that, he was a Senior Reporter at TechCrunch for 13 years. He wrote over 3,500 articles on technology and tech startups and established himself as an influential voice on the European tech scene.

AI and Privacy: Building secure intelligence at scale

Bart is one of the first employees of Proton and has been CTO since 2015. He is responsible for technical vision and architecture. Previously, he worked at the Large Hadron Collider at CERN and in the private sector. He holds a BA in physics and chemistry from Dartmouth College, a PhD in physics from Stanford University and was a postdoctoral fellow at Harvard University.

Jeremie Eliahou Ontiveros leads Datacenter & Energy Infrastructure research at SemiAnalysis. He specializes in tracking Hyperscaler and Neocloud infrastructure and large-scale AI clusters, cooling & electrical systems, and power sources. He connects systems-level architecture with real-world deployment to decode AI’s backend complexity. Previously, he focused on long-only investments in public equities, where he honed his expertise in deep data analysis and financial modeling.

Jeremie holds a Bachelor’s degree from University of Lille 1 Sciences and Technology, as well as a Master’s degree from IAE Gustave Eiffel.

Building at the speed of agents: from training pipelines to data infrastructure

Dr. Antoine Jardin serves as Chief Technology Officer of Arlequin AI, bringing exceptional expertise in algorithmic and statistical sciences to the role. An accomplished engineer and data scientist from CNRS (French National Centre for Scientific Research), he holds a doctorate in algorithmic and statistical sciences.

His influence extends to the national level, having contributed to the definition of France's 2018 National AI Strategy and the implementation of the Jean Zay supercomputer, one of France's most powerful computing resources. These contributions have helped shape the country's artificial intelligence infrastructure and strategic direction.

Prior to his current position, he served as Technical Director of the Radicom project, a notable collaboration between Princeton University and the World Bank.

Inductive learning, AI safety, and superintelligence in the age of disinformation

Steeve Morin is the Founder and CEO of ZML, a next-generation AI inference engine built to deliver peak performance across a wide range of hardware platforms. Previously, he served as VP of Engineering at Zenly (acquired by Snap Inc.), where he led engineering efforts for millions of users. With a deep software engineering background and expertise in high-performance systems, Morin now drives ZML to push the boundaries of inference performance and developer experience.

Inference Everywhere: optimizing performance

A former neuroscience Ph.D. student turned tech leader, Alexandre fell in love with software engineering for its mix of logic, creativity, and impact. Today, as VP of Engineering at Aive, he drives the company's technical vision and leads the teams behind its AI-powered video creation platform.

After joining Aive as Senior Backend Engineer, he built and scaled the backend team from scratch, redesigned the architecture, and strengthened collaboration between product and data science. His leadership blends analytical precision with a people-first mindset, fostering clarity, autonomy, and innovation across teams.

A black belt in judo with over 16 years of practice, Alexandre draws on discipline and balance to lead with focus and resilience. Guided by his motto, "You can go a long way with a smile," he believes that technology is as much about human connection as it is about innovation.

Decoding Aive's technology behind next-gen video transcreation: MGT

Dam Mulhem graduated from the Institut Supérieur des Techniques du Son and began his career in the media industry as a Sound and Image Engineer. He then joined Apple, where he took on the role of Training Manager.

In 2015, Dam co-founded XXII alongside William Eldin, taking on the position of Product Director, where he has been shaping innovative solutions at the intersection of technology and creativity.

Olivier Reynaud has been exploring the creative infrastructure of tomorrow for over a decade.

After co-founding Teads (which became the global leader in programmatic video), he co-founded Aive in 2019 alongside Rudy Lellouche (ex-Adyoulike) with a radical thesis: creativity is no longer a handcrafted process — it’s an augmented, measurable, and scalable force. With MGT (Multimodal Generative Technology), Aive merges analytical and generative AI into a single pipeline — recreating content that preserves its original intent while adapting seamlessly to different contexts. This is transcreation, not versioning — the antithesis of closed, hallucinatory models.

A creative mind obsessed with execution, Olivier champions a responsible, European-born AI that is GDPR-native, AI Act compliant, and relies on zero external APIs.

Aive has raised over €25 million since its inception and now serves dozens of clients across France and the U.S., including L’Oréal, Stellantis, P&G, TF1, Publicis, IPG, and Meta.

His conviction: the future of video won’t just be generated — it will be transcreated.

That’s what he calls Creative Intelligence.

Gaël Delalleau is the founder and CEO of Kog, a European VC-funded startup and real-time AI frontier lab. With a world-class team of 15 engineers and PhDs, he is unlocking the future of AI use cases by building optimized inference engines and training new models architectures that achieve 100x faster token generation.

With a strong engineering and research background as an École Polytechnique graduate, he previously built a handful of web, gaming, optimization, machine leaning and cybersecurity products at several startups and consulting companies.

Real-time AI: building 100x faster inference

Dr. Corina Chutaux is a researcher and lecturer collaborating with Sorbonne University, where she contributes to projects at the intersection of the human sciences and artificial intelligence. She holds a PhD from Sorbonne University, with research focusing on the relationship between creativity and algorithmic generation, and on how machine learning reshapes artistic and literary practices. Formerly a lecturer at Sorbonne Nouvelle and Sciences Po, Dr. Chutaux has developed an interdisciplinary approach bridging aesthetics, digital culture, and computational logic. She is the author of Esthétique de l’Art Invisuel and a contributor to a collective volume featuring her chapter Art and Literature through Generative Adversarial Networks. She has also contributed to scientific papers on natural language processing (NLP), particularly in the design and interpretation of linguistic corpora. Her research examines the epistemological and ethical implications of AI-driven creativity within a critical humanistic framework.

Creative Integrity and Algorithmic Innovation: A Dual Approach

Co-Founder of BLACKBOX AI. Building Full Self Coding to transform how developers write code. On a mission to make programming accessible to everyone. 

Agents that actually do the work: how autonomy changes the way we build

Alexandre Pasquiou co-founded Neuralk-AI, a deeptech startup based at Station F (Future 40) developing verticalized Tabular Foundation Models for enterprise tabular data and use cases. Alexandre holds an engineering degree from CentraleSupélec and a PhD in computational neuroscience from Inria, where he conducted research in the MIND team under the supervision of Bertrand Thirion and Christophe Pallier.

Tabular Foundation Models: the new frontier of data science

Pauline Pham oversees strategy and operations at Dust, a Sequoia-backed AI startup building workplace AI agents from SF and Paris. Companies like Cursor, Vanta, Clay, Doctolib, Mirakl, and Qonto build with Dust. Before Dust, Pauline worked in operational roles at fast-growing technology and strategy companies. She is passionate about building impactful products and empowering everyone with the power of AI.

From single agents to agent fleets: lessons from building Dust's agentic platform

I'm a Solutions Architect with a focus on LLM-based applications. I lead the AI Team at SOFTIQ, turning ideas into real AI solutions. I work hands-on with clients to design products that meet their needs and deliver quick Proof of Concepts to get things moving fast. My experience includes RAG pipelines, agent systems, and working with open source language models, including those for underrepresented languages like Polish.

Building computer-using agents for AI lead generation

After having spent more than 20 years in various automotive R&D positions managing powertrain and vehicle projects, Frédéric Petit is now Research and Innovation Director in charge of developing Beyond Automotive activities for Valeo Power Division.

In this role, he is leading Datacenter innovation projects focusing on technologies such as 2-Phase Direct-To-Chip and Immersion Cooling.

Dr Frédéric Petit owns a PhD from Louvain University, an Engineering Master degree from Mons Polytechnic School (Belgium) and an EMBA from HEC Paris.

Andrei Bursuc is a Deputy Scientific Director and Senior Research Scientist at valeo.ai (the AI upstream research lab of  the Valeo group), as well as an associate researcher at Inria (National Institute for Research in Digital Science and Technology) Paris, within the Astra project team. He works on perception for assisted and autonomous driving. His research focuses primarily on the reliability of AI models (deep neural networks), learning with limited supervision, and multimodal and multi-sensor perception. He also teaches at École Polytechnique and École Normale Supérieure in Paris.

Andrei obtained his engineering degree from Politehnica University of Bucharest in 2009. After a Ph.D. at École des Mines de Paris and Alcatel-Lucent Bell Labs France in 2012, he was a postdoctoral researcher at Inria Paris and Rennes, then a researcher at Safran Tech in aeronautics. A member of the ELLIS society, he regularly contributes to the program committees of major international scientific conferences in artificial intelligence and computer vision: CVPR, ICCV, ECCV, and NeurIPS, and has co-organized and presented several tutorials on self-supervised learning as well as on reliability and uncertainty estimation at these conferences.

Open & Repurposable: Foundation Models for the Automotive Industry

Alina Holcroft, CEO and founder of ETHIQAIS, is an internationally recognized expert in ethical AI. A member of the France 2030 Digital Committee and the AI ​​standardization committee at AFNOR, ISO, and CEN/CENELEC, she holds the IEEE AI system auditing accreditations currently available in France. Recently awarded the Tech for Change prize at Vivatech 2025 and the Start 2025 trophy from Réseau Entreprendre Alsace, Alina Holcroft was also graded among the top 5 entrepreneurs in Alsace by the Choiseul 2025 ranking

AI Adoption: Measuring and Maximizing ROI within the European Regulatory Framework

Enrico Bertino is the Chief AI Officer of indigo.ai, a leading European player in generative AI for natural language. With a background in Mathematical Engineering from the Polytechnic University of Milan and expertise in Big Data and machine learning, he co-founded Indigo.ai to drive innovation in conversational AI. He leads the company’s R&D and customer success team, ensuring cutting-edge language models like BERTino, one of the most downloaded Italian NLP models, are effectively developed and implemented. A Forbes Under 30 honoree, he collaborates with top universities and research institutes to advance AI and its real-world applications.

Benjamin de Almeida is the CEO and founder of SOCLE AI, a platform reinventing how European developers and organizations deploy artificial intelligence: sovereign, compliant, and uncompromising on user experience.

Before launching SOCLE AI, he served as CPO at Golem.ai, where he led the development of explainable AI solutions for major public and private clients.

An entrepreneur and passionate builder, Benjamin promotes a pragmatic vision of AI, where sovereignty has its place, without dogmatism. He regularly speaks on topics such as technological sovereignty, compliance (GDPR, AI Act), transparancy, trust, and the transformation of business practices through AI.

Antoine de Torcy, Co-founder & Chief AI Scientist, directs BIOLEVATE's core technology. He applies his expertise in industrial AI, developed at Dataiku and Dassault Systèmes, to accelerate discovery in the life sciences.

Joël Belafa is a passionate technologist and the Co-founder and CEO of Biolevate, a Paris-based startup leveraging artificial intelligence to streamline medical writing, regulatory documentation and drug-development workflows. He served in executive engineering roles at Dataiku (including Director of Engineering) where he helped build out enterprise-AI platforms for large organisations. Beyond his executive roles, he acts as an advisor to multiple technology startups and scale-ups, sharing his expertise in AI, cloud computing and operationalising data science. His contributions span both strategic leadership and technical execution in the intersection of DataOps and AI.

Cas Burggraaf is the Chief Technology Officer and co‑founder of GreenPT, a Dutch AI startup that delivers sustainable and privacy‑friendly artificial intelligence. With a decade of software development experience, he leads the technical vision, steering the deployment of open‑source, open‑weight models. 

How to build a sustainable AI infrastructure

Sophia is the founder of Biostream, a start-up developing wearables that use AI-based sensor fusion to detect critical emergencies in real time. The company operates across defense, emergency response, and high-risk civilian environments.

Prior to Biostream, she founded and exited Meltdown, a global gaming franchise. She holds an MBA from the University of Oxford and works as an advisor at Arma Partners.

Anis is Principal AI at Veepee, where he drives company-wide adoption of Generative AI, from strategy and roadmap design to the delivery of end-to-end AI products. He advises executive, technical, and business teams on how to turn emerging AI capabilities into measurable impact.

Previously, Anis co-founded and led daco, a retail AI startup that used proprietary algorithms to benchmark competitors’ assortments, pricing, and discount strategies; daco was acquired by Veepee in 2018. Within Veepee, he then built the group’s pricing toolchain and led product and engineering teams on strategic data and AI initiatives, including large-scale e-commerce scraping, product data enrichment, and pricing optimization across Europe.

Cracking Product taxonomy at scale: A multimodal, zero-shot approach

David Bertoin is a research engineer at Photoroom, where he focuses on improving generative models for creative image editing and leading the PRX open-source text-to-image project. He also teaches as a part-time associate professor at INSA Toulouse. David’s work centers on representation learning, spanning computer vision and reinforcement learning. Before joining Photoroom, he was a research scientist at IRT Saint-Exupéry, contributing to the DEEL and ANITI research programs.

Inside Photoroom’s Open-Source T2I model

Jean-Baptiste Kempf is the Chief Technology Officer at Scaleway, a leading European cloud infrastructure provider. With over 20 years of experience in open source software—most notably as a major contributor to the VLC media-player and the VideoLAN project—he brings deep engineering expertise and a community mindset to his role. At Scaleway he leads the technical strategy and drives innovation across cloud services, while maintaining a strong commitment to open-source principles.

From Foundation Models to Real-World Actions

I'm a Product Manager at Scaleway and lead the development of their environmental footprint calculator. This calculator provides real-world usage metrics for cloud infrastructure CO₂ emissions and water consumption. With an engineering degree and technical experience as a former developer, I bring a methodical and practical approach to creating tools that help developers and DevOps teams make informed and sustainable decisions.

Driven by curiosity and sustainable development, my work aims to provide transparent and actionable information to cloud users.

Build the OCR-powered menu inventory for “Café de l’AI”

Created Taveo, an electronics design house, ran it for 6 years, then did some pretty cool stuff at Scaleway for 14 years.

Built 12 products there as a "hands-on" Engineering Manager. Designed servers, routers, firmwares, BIOSes, drivers, embedded software, backends, APIs, ... you name it!

Build the audio agent for “Café de l’AI”

Nathalie Birocheau is one of the leading women shaping the future of music and technology. An engineer from Centrale-Supélec and former strategy consultant, she has led major cultural and media projects, including the launch of franceinfo as Deputy Director of France Info. Six years ago, she co-founded Ircam amplify, the commercial arm of IRCAM, bringing sound innovation to the heart of creative industries. 

Han HELOIR YAN, Ph.D., is passionate about AI system architecture and how it concretely transforms businesses. At Mistral AI, as a Partner Solutions Architect covering EMEA, she helps make generative AI truly adoptable for enterprises. With 15 years of experience in finance (Murex), telecommunications (Orange, Huawei), and databases (MongoDB), she shares in-depth insights on multi-agent systems, multimodal AI, and production architectures through her articles. Her approach is pragmatic, analytical, and always driven by curiosity to explore what truly works beyond the hype. She believes in elegant, scalable solutions and the importance of sharing high-quality technical knowledge.

Agentic Stack for Regulated Industries: Architecture Essentials

Frederic Amadu is the CTO of Ircam Amplify with over 20 years of experience in audio DSP and AI. An engineer and inventor with several patents, he has led the development of advanced signal-processing technologies and AI-driven audio systems. He focuses on turning cutting-edge research into scalable innovative audio products.

Develop business in AI and HPC data center, focusing on advanced architectures for NVIDIA HGX and GB NVL72, AMD Instinct, and Intel Gaudi platforms. My experience includes CDU, direct liquid cooling, immersion cooling, GPU/CPU system design, HPC cluster development, and end-to-end data center implementation. Support AI/HPC tenders and large-scale deployments across academia, enterprises, and cloud providers, covering cooling integration, network architecture, sustainability requirements, and system delivery. With cross-regional and cross-functional project experience, I help customers build scalable, efficient, and future-ready data centers.

Firas Abi-Farraj is the CTO of AI and robotic manipulation at Enchanted Tools, where he works on bringing robots out of the lab and into the real world. He has worked at several leading European robotics institutes, including ETH Zurich, DLR, Inria, and Engie, focusing on robotic systems for human assistance across domains such as care, space exploration, and the nuclear industry. He holds a PhD in robotics from Inria and was a postdoctoral fellow at ETH Zurich.

Sean Varley is the Chief Evangelist and leader of ecosystem development and partner programs at Ampere Computing.  His group is responsible for go-to-market, messaging, strategy and execution of developer, partner and revenue ecosystem growth programs. The team covers marketing strategy, strategic business relationships, business planning and developer recruitment for AI Computing in the rapidly evolving Cloud and Edge server markets.   At Ampere, our goal is to provide Products and Solutions that showcase the disruptive advantage our unique processors provide to AI and traditional computing markets. We elegantly integrate Cloud Native SW and HW that help our customers understand and use our products to drastically transform their AI computing investments into highly efficient, high performance, imminently sustainable service infrastructures.

Building Al that scales: efficient compute for a smarter, faster, everywhere world 

As co-CTO at Enchanted Tools, I lead the technical strategy to create a new generation of iconic robots, combining world-class engineering with character design to bring technology to life.

My background is deeply rooted in scaling complex autonomous systems. Previously, as co-founder and CTO at Azur Drones, I led the development of the first certified industrial Drone-in-a-box system in Europe. Today, I bridge the gap between deep-tech mechatronics and high-level AI strategy, driven by a passion for exploring how robotics can reshape our society.

As Chief Operating Officer at Scaleway, Albane oversees the company’s global operations, including customer care, the supply and hardware chain across its 12 data centers in Europe, and all strategic operational projects. She drives major transformations to position Operations, Hardware, Customer Care, and Finance as key enablers of innovation and international growth.

Before joining Scaleway, Albane built her expertise at L'Oréal and LVMH, where she developed a strong sense of rigor and excellence. She is recognized as an inspiring leader, able to blend strategic vision, financial insight, and a collaborative approach.

A graduate of Kedge Business School, she leads by example and supports others as a mentor and advisor, notably through initiatives like Startup Weekend Women. At Scaleway, she also drives CSR projects aligned with the company’s commitment to responsible innovation.

Aurélien Delfosse is a Core-Team Researcher at H Company, based in the Paris area. With a solid engineering foundation, he brings deep expertise in machine learning, algorithmic design, and large-scale AI systems.

Early in his career, he worked on reinforcement learning applied to physics-driven problems, developing optimization and control methods in complex simulated environments.

He later joined InstaDeep, where he played a key role in scaling the research team and building applied-research capabilities across projects and disciplines, like biology, logistic and defence. His work spanned both technical leadership and hands-on research, contributing to the growth of teams and production-grade ML systems.

Currently, he focuses on scaling model teams and infrastructure at H company with an emphasis on performance, reliability, and real-world impact with agentic AI.

With a degree in Electronic Engineering and as the lead inventor on a number of patents, Dan has worked across a range of industries - semiconductor design, software defined radio, data centre infrastructure and data storage.

As VAST Data’s lead for AI service providers (“CSPs”), Dan works with organisations building out next-generation data-centric platforms and services

From Desktop to Supercomputer: Building Local AI Workflows with GIGABYTE AI TOP ATOM

Robert Keus is a Dutch entrepreneur and AI strategist with a mission to make technology more responsible, accessible, and sustainable. As founder of Brthrs Agency and GreenPT.ai, he blends nearly two decades of digital innovation experience with a commitment to social impact. Robert has worked with organizations ranging from public institutions to global brands, always with a focus on ethical AI, privacy compliance, and inclusive design. His vision for GreenPT is to prove that AI can be both cutting-edge and climate-conscious - helping businesses harness its potential without compromising on ethics or the environment.

Fred Bardolle leads AI Product at Scaleway, where he drives the vision and delivery of Europe’s sovereign, efficient, and open AI infrastructure. He oversees inference-as-a-service products, large-scale GPU training clusters, and the AI tooling that helps make Scaleway an AI-first company.

He is the host of C’est pas SorciAI, a show produced with Le Ticket, where he explores AI topics with a clear, accessible tone. Fred also speaks regularly at conferences and events, sharing his vision for an AI ecosystem with a strong European soul.

Before Scaleway, he worked across data science, product, and engineering-heavy environments, bringing a clear, curious, and hands-on approach to complex systems. He previously founded and led France’s equivalent of the Defense Digital Service, bringing modern software practices to national security.

Fred holds a Ph.D. in machine learning and an engineering background. AI moves fast. To stay sharp, he practices meditation.

Henrik Landgren is the Co-founder and CPTO at Gilion, an AI-powered funding platform that has deployed €100M+ with zero credit losses. Henrik was previously Spotify's first VP of Analytics, where he led data strategy under Daniel Ek throughout the company's explosive growth to over 20 million subscribers and 2,000 employees. He then built EQT Ventures' Funds I and II from launch to over 100 investments, while simultaneously creating Motherbrain, an AI-driven investment platform that revolutionised data-driven deal sourcing and decision-making. At Gilion, he leads development for the firm’s AI-powered investment platform. By analysing 1 billion+ live data points per company, Gilion's platform predicts revenue 12 months ahead with 90%+ accuracy. Gilion is also offering its technology as a product, enabling investors to use agentic AI to get greater granularity on their investment data. 

Finally solving the high-stakes problems with agentic tree structures and metrics pipes

Bruno Lecointe is the Group Vice President of HPC, AI, and Quantum Business at Eviden, where he focuses on aligning offerings with customer objectives. His career has been closely tied to advancements in scientific simulation technology. Bruno leads by driving innovation and fostering collaboration in data security, advanced analytics, and digital resilience.

He began his career as a Business Developer at BONATRE Frères and later advanced to roles such as Sales Director at ALINKA, where he promoted Linux cluster software in HPC.

At NEC Europe, he excelled as Senior Sales Manager and Managing & Operations Director, driving sales growth and team expansion.

Bruno also serves as the Treasurer of Teratec, reinforcing his commitment to advancing high-performance computing and technological excellence across the industry.

Born in Tokyo and raised in Paris, Romain Takeo Bouyer has spent the past decade in London and New York contributing to the growth of Spotify, the global leader in audio streaming. He leads a data science group working across music, podcasts, and audiobooks, guiding business and strategy decisions. His teams analyze Spotify’s partnerships with rights holders, drive insights that help refine Spotify’s recommendations and policies in the age of AI. 

Romain previously worked at Criteo, focusing on media buying via machine learning, and at Deezer, a music-streaming pioneer. He recorded artists across four continents, composed a musical, founded an advisory committee for the Palais de Tokyo, and serves on the board of the Château du Clos Lucé, Leonardo da Vinci’s last home. He lectures regularly at NYU, ESSEC and MIT.

AI everywhere: reinventing Spotify's core strategies

Ramine Roane is the Corporate Vice President of AI Product Management and ecosystem development at AMD, based in San Jose, California. Prior to this role, he served as Vice President of Data Center Acceleration within AMD’s Adaptive and Embedded Computing Group in 2022. Before the AMD-Xilinx merger, Ramine was Vice President of Software and AI Product Management at Xilinx, where he oversaw Electronic Design Automation (EDA) software tools. Earlier in his career, he held leadership roles at EDA companies including Synopsys and Mentor Graphics. Ramine holds a Master’s degree in Electronic Engineering and Computer Science from the National Polytechnic Institute of Grenoble in France

He is a research scientist at Photoroom, focusing on innovative projects in advanced AI, including generative models. Before this, he worked as a research scientist at Naver Labs Europe and as a research lead at V7 Labs.

His academic background includes a Ph.D. in Computer Vision from the Universitat Autònoma de Barcelona and a postdoctoral position at Xerox under the supervision of Florent Perronnin. His research interests lie in computer vision and machine learning, with a current focus on representation learning, multi-modal generative models, and self-supervised problems.

Constance Morales is a Senior Product Marketing Manager at Scaleway, where she has spent the past three years leading the go-to-market strategy for the company’s entire AI portfolio and a broad range of Cloud products, including cybersecurity, environmental impact, FinOps and marketplace offerings.

With a deep understanding of the Cloud and AI markets, Constance has played a central role in shaping Scaleway’s position in the European AI ecosystem. She led the GTM initiative behind Scaleway’s supercomputers in 2023, a milestone that helped redefine the company’s strategy and market presence.

Since the very first edition, she has also overseen the editorial direction of ai-PULSE, Scaleway’s annual AI event gathering 1,500 attendees and more than 50 industry experts each year. She built the event’s editorial engine from scratch and has refined it across three consecutive editions.