01
LLM training
Pretraining data, scaling laws, synthetic data, RL from execution feedback, distillation and post-training systems built for production models.
3rd Paris edition · Registration now open
GenerationAI
The “NeurIPS” for AI engineers. 3rd edition.
The conference for AI management from training to production, and the agent engineering discipline that turns frontier models into dependable systems. The 3rd GenerationAI in Paris, and the 18th worldwide.
Location
CNIT Forest, Paris
Date
December 1-2, 2026
Edition
3rd Paris · 18th worldwide
3
Paris edition
18
worldwide
80+
speakers
2000+
attendees
8
research tracks
3rd Paris edition // Applied AI
Manage AI from the first training run to every decision in production.
For its 3rd Paris edition, GenerationAI connects model research to the systems discipline required to operate intelligent products in the real world.
01
Train the model
02
Engineer the agent
03
Operate the system
04
Measure the outcome
3rd Paris edition // Featured speakers
The researchers and builders defining the next generation of AI.
From open-weight robotics and web-scale crawling to AI governance, HPC infrastructure, and production systems.
Hugo Duminil-Copin
Fields Medalist
IHES
Simon Frieder
AIMO (AI Mathematical Olympiad) Prize Manager
AIMO Prize
Remi Cadene, PhD
CEO
Steven Palma Morera
Tech Lead LeRobot
Thom Vaughan
Principal Engineer
Simen Yang
Member of Technical Staff
OpenAI
Paul Emile Brotons
Member of Technical Staff, Applied AI
Anthropic
Hector Gonzales
CEO
spiNNcloud
Jeremie Doucy
CTO
Bayes Impact
Damien Wattez
AI and Innovation Director
SNCF
Victor Schmidt
CEO
Entalpic
Gilles Closet
Global AI Ecosystem Leader
Matthieu Boussard
Co-Founder & VP AI
Craft.AI
Stéphie Herlin
Co-Founder
KORA
Alexandre Gefen
Research Director, author of "A Cultural History of AI"
CNRS
Neil Zeghidour, PhD
CEO
Djalal Benbouzid, PhD
Senior AI Expert
Olivier Debeugny
CEO
Julie Hunter, PhD
Senior Researcher
Stephane Requena
CTO
Jean-Baptiste Kempf
Creator of VLC, CEO
Steeve Morin
CEO
Anastasia Stasenko, PhD
CEO
Marie Paindavoine
CEO
Skyld
Yann Jouanin
CEO and Co-Founder
MCP Apps Builders
Frédéric Barthelet
Co-founder and CTO
Alpic
Romain Dillet
Author and tech journalist
Laurie Soffiati
CEO France
Nabla
Conference theme
Open weight models vs Frontier Model
Defining the future of API
Will applications call closed frontier APIs, or run open-weight models behind their own endpoints? GenerationAI puts that defining architectural choice at the center of every track, talk and hallway conversation.
Agent engineering core
Agent engineering is the new production engineering.
The experience people supervise. The harness connecting tools, memory and context. The loops that plan, recover and finish the job.
Agent experience
Harness engineering
Loop engineering
Who is in the room
Built for people who ship models.
- +AI engineers
- +AI researchers
- +AI-native founders
- +AI tech decision makers
Every session is presented by the team that produced the result, with the methodology on the slide.
3rd Paris edition // 8 tracks in parallel
The complete AI lifecycle. From training clusters to production agents.
02
Agent experience
Designing the human surface of autonomous work: approvals, interruption, legibility, memory, trust calibration and collaborative control.
03
Harness engineering
Tool design, sandboxing, permissions, context assembly, observability and the execution layer that turns a model into a reliable system.
04
Loop engineering
Control flow for non-deterministic workers: planning, retries, reflection, context compaction, recovery and long-running agent loops.
05
Open weights vs frontier
What open-weight releases can now do, where closed frontier models still lead, and how that choice defines the future of the API.
06
Benchmarks & evals
Contamination, saturation, agentic task suites and internal evaluations that remain useful through model and harness changes.
07
Inference systems
Serving economics, KV cache strategy, speculative decoding and latency budgets for multi-step, high-concurrency agent workloads.
08
MCP & interoperability
Model Context Protocol servers in production: capability negotiation, authentication and the emerging interoperable tool ecosystem.
Agent systems
Design the loop, not just the prompt.
LLM training
Train the model for the real world.
80+ speakers
From frontier labs to the teams running models at massive production scale.
Meet researchers building the models, scale-up leaders turning them into fast-growing products, and engineering teams from large enterprises operating AI across complex, high-volume production systems.
OpenAI
Anthropic
Google Gemini
Microsoft Copilot
Mistral
Google DeepMind
HuggingFace
Nvidia
AMD
Intel
Kyutai
Gradio
ElevenLabs
Gladia
Pleias
INRIA
Fraunhofer Institute
Meta AI
Cohere
AI2
Nous Research
Allen Institute for AI
Stability AI
+ many more
Past speakers // The road to edition 03
Researchers and builders who already took our stage.
From frontier labs and open-weight teams to enterprises running AI in production.
Michal Valko
ex-Principal Engineer, Llama team
FAIR, Meta
Thomas Mesnard
Research Scientist (ex-Google DeepMind)
Meta Superintelligence Labs
Andrés Marafioti
Research Lead, Multimodal Models
Hugging Face
Yanis Tevissen
Head of AI Research
MomentsLab
Loubna Ben Allal
Researcher, SmolLM Lead
Hugging Face
Mehdi Ghissassi
CTO (ex-Google DeepMind)
AI71
Djalal Benbouzid, PhD
AI Governance
Volkswagen
Vincent Gonguet
Head of Product, GenAI Llama Safety
Meta
Sai Prasad Veluru
AI Engineer
Apple
Rahul Raja
Staff Software AI Engineer
LinkedIn
Daniel Vahos
AI Engineer
Valeo
Lê Nguyên Hoang
Researcher & Professor, AI Cybersecurity, author and YouTuber
Tournesol
Wassim Bouaziz, PhD
AI Researcher
Meta
Lelio Renard
Head of Engineering
Mistral
Romain Huet
Head of Developer Experience
OpenAI
Neil Zeghidour
CEO
Gradium
Jean-Baptiste Kempf
Creator of VLC, CEO
Kyber
Steeve Morin
CEO
ZML
Axelle Arquié
AI and Labor Market Economist
CEPII
Gabriel Hubert
CEO
Dust
Anastasia Stasenko, PhD
CEO
Pleias
Jean-Gabriel Barthélemy
AI Researcher
DragonLLM
Thibaut Veron
Chief of Staff
French Center for AI Safety (CeSIA)
Damien Wattez
Head of AI, TGV
SNCF Voyageurs
Ziv Ilan
AI Startups
NVIDIA
Yann Lechelles
CEO
Probabl
Élisabeth Mouchy
Lead PM AI
Betclic
Past sponsors
The companies that powered our previous editions.
From chip makers and cloud platforms to consultancies and applied AI startups: the partners who supported the Paris 2025 edition.
%20--%3e%3csvg%20version='1.1'%20id='IBM_logo_8-bar_blue'%20xmlns='http://www.w3.org/2000/svg'%20xmlns:xlink='http://www.w3.org/1999/xlink'%20x='0px'%20y='0px'%20width='1075px'%20height='401.15px'%20viewBox='0%200%201075%20401.15'%20enable-background='new%200%200%201000%20401.15'%20xml:space='preserve'%3e%3ctitle%3eLogo%20International%20Business%20Machines%20Corporation%3c/title%3e%3cg%20id='IBM_logo'%20fill='%231f70c1'%3e%3cg%20id='I'%3e%3crect%20y='373.17'%20width='194.43'%20height='27.932'/%3e%3crect%20y='319.83'%20width='194.43'%20height='27.932'/%3e%3crect%20x='55.468'%20y='266.54'%20width='83.399'%20height='27.932'/%3e%3crect%20x='55.468'%20y='213.25'%20width='83.399'%20height='27.932'/%3e%3crect%20x='55.468'%20y='159.96'%20width='83.399'%20height='27.932'/%3e%3crect%20x='55.468'%20y='106.58'%20width='83.399'%20height='27.932'/%3e%3crect%20y='53.288'%20width='194.43'%20height='27.932'/%3e%3crect%20width='194.43'%20height='27.932'/%3e%3c/g%3e%3cg%20id='B'%3e%3cpath%20d='m222.17%20400.85%20207.11%200.297c27.734%200%2052.793-10.697%2071.513-27.932h-278.62z'/%3e%3cpath%20d='m222.17%20347.76h299.03c5.051-8.617%208.815-18.027%2011.094-27.932h-310.12z'/%3e%3crect%20x='277.73'%20y='266.54'%20width='83.3'%20height='27.932'/%3e%3cpath%20d='m444.43%20266.54v27.932h90.927c0-9.608-1.288-19.017-3.764-27.932z'/%3e%3cpath%20d='m497.92%20213.25h-220.19v27.932h243.46c-6.34-10.698-14.165-20.107-23.277-27.932z'/%3e%3cpath%20d='m277.73%20159.96v27.932h220.19c9.311-7.825%2017.135-17.235%2023.277-27.932z'/%3e%3crect%20x='277.73'%20y='106.58'%20width='83.3'%20height='27.932'/%3e%3cpath%20d='m444.43%20134.51h87.163c2.476-8.914%203.764-18.324%203.764-27.932h-90.927z'/%3e%3cpath%20d='m521.2%2053.288h-299.03v27.932h310.12c-2.575-9.905-6.339-19.314-11.093-27.932z'/%3e%3cpath%20d='m429.28%200h-207.11v27.932h278.53c-18.621-17.235-43.878-27.932-71.414-27.932z'/%3e%3c/g%3e%3cg%20id='M'%3e%3cpolygon%20points='555.57%2081.22%20742.67%2081.22%20733.06%2053.288%20555.57%2053.288'/%3e%3cpolygon%20points='555.57%2027.932%20724.25%2027.932%20714.64%200%20555.57%200'/%3e%3cpolygon%20points='861.03%20401.17%20861.03%20373.24%201e3%20373.24%201e3%20401.17'%20stroke-width='1.0018'/%3e%3cpolygon%20points='861.03%20347.76%20861.03%20319.83%201e3%20319.83%201e3%20347.76'/%3e%3cpolygon%20points='777.73%20182.54%20769.91%20159.96%20694.43%20159.96%20611.03%20159.96%20611.03%20187.89%20694.43%20187.89%20694.43%20162.24%20703.25%20187.89%20852.22%20187.89%20861.03%20162.24%20861.03%20187.89%20944.43%20187.89%20944.43%20159.96%20861.03%20159.96%20785.56%20159.96'/%3e%3cpolygon%20points='944.43%20106.58%20803.98%20106.58%20794.37%20134.51%20944.43%20134.51'/%3e%3cpolygon%20points='1e3%2027.932%201e3%200%20840.93%200%20831.32%2027.932'/%3e%3cpolygon%20points='768.13%20373.22%20777.73%20400.85%20787.34%20373.22'/%3e%3cpolygon%20points='749.5%20319.83%20759.31%20347.76%20796.16%20347.76%20806.06%20319.83'/%3e%3cpolygon%20points='730.78%20266.54%20740.59%20294.47%20814.88%20294.47%20824.68%20266.54'/%3e%3cpolygon%20points='721.97%20241.18%20833.6%20241.18%20843.11%20213.25%20712.36%20213.25'/%3e%3cpolygon%20points='611.03%20134.51%20761.09%20134.51%20751.49%20106.58%20611.03%20106.58'/%3e%3cpolygon%20points='1e3%2053.288%20822.4%2053.288%20812.9%2081.22%201e3%2081.22'/%3e%3crect%20x='555.57'%20y='373.22'%20width='138.97'%20height='27.932'/%3e%3crect%20x='555.57'%20y='319.83'%20width='138.97'%20height='27.932'/%3e%3crect%20x='611.03'%20y='266.54'%20width='83.399'%20height='27.932'/%3e%3crect%20x='611.03'%20y='213.25'%20width='83.399'%20height='27.932'/%3e%3crect%20x='861.03'%20y='213.25'%20width='83.399'%20height='27.932'/%3e%3crect%20x='861.03'%20y='266.54'%20width='83.399'%20height='27.932'/%3e%3c/g%3e%3cpath%20id='Registered'%20d='m1052%20357.15a22%2022%200%200%200-22%2022%2022%2022%200%200%200%2022%2022%2022%2022%200%200%200%2022-22%2022%2022%200%200%200-22-22zm0%204a18%2018%200%200%201%2018%2018%2018%2018%200%200%201-18%2018%2018%2018%200%200%201-18-18%2018%2018%200%200%201%2018-18zm-9.3476%206.793v22.414h5.453v-7.7305h3.0978l4.164%207.7305h5.9804l-5.0234-8.582c2.4616-0.96446%204.0624-3.2158%204.0624-6.75%200-4.0818-2.5594-7.082-7.582-7.082zm5.453%204.2891h4.0548c1.7114%200%202.668%200.75016%202.668%202.3555v1.6133c0%201.6053-0.9566%202.3594-2.668%202.3594h-4.0548z'/%3e%3c/g%3e%3c/svg%3e)
%20--%3e%3csvg%20version='1.1'%20id='Layer_1'%20xmlns='http://www.w3.org/2000/svg'%20xmlns:xlink='http://www.w3.org/1999/xlink'%20x='0px'%20y='0px'%20viewBox='0%200%20388.2%20150.6'%20style='enable-background:new%200%200%20388.2%20150.6;'%20xml:space='preserve'%3e%3cstyle%20type='text/css'%3e%20.st0{fill:%2304C7FD;}%20.st1{fill:%230068B5;}%20%3c/style%3e%3crect%20y='2.1'%20class='st0'%20width='28.1'%20height='28.1'/%3e%3cpath%20class='st1'%20d='M27.4,148.5V47.3H0.8v101.2H27.4z%20M204.2,149.5v-24.8c-3.9,0-7.2-0.2-9.6-0.6c-2.8-0.4-4.9-1.4-6.3-2.8%20c-1.4-1.4-2.3-3.4-2.8-6c-0.4-2.5-0.6-5.8-0.6-9.8V70.1h19.3V47.3h-19.3V7.8h-26.7v97.9c0,8.3,0.7,15.3,2.1,20.9%20c1.4,5.5,3.8,10,7.1,13.4s7.7,5.8,13,7.3c5.4,1.5,12.2,2.2,20.3,2.2L204.2,149.5L204.2,149.5z%20M357,148.5V0h-26.7v148.5H357z%20M132.5,57.2c-7.4-8-17.8-12-31-12c-6.4,0-12.2,1.3-17.5,3.9c-5.2,2.6-9.7,6.2-13.2,10.8l-1.5,1.9v-1.7V47.3H43v101.2h26.5V94.6v3.7%20c0-0.6,0-1.2,0-1.8c0.3-9.5,2.6-16.5,7-21c4.7-4.8,10.4-7.2,16.9-7.2c7.7,0,13.6,2.4,17.5,7c3.8,4.6,5.8,11.1,5.8,19.4l0,0v0.2l0,0%20l0,0v53.5h26.9V91C143.7,76.6,139.9,65.2,132.5,57.2z%20M316.5,97.7c0-7.3-1.3-14.1-3.8-20.5c-2.6-6.3-6.2-11.9-10.7-16.7%20c-4.6-4.8-10.1-8.5-16.5-11.2s-13.5-4-21.2-4c-7.3,0-14.2,1.4-20.6,4.1c-6.4,2.8-12,6.5-16.7,11.2s-8.5,10.3-11.2,16.7%20c-2.8,6.4-4.1,13.3-4.1,20.6c0,7.3,1.3,14.2,3.9,20.6c2.6,6.4,6.3,12,10.9,16.7c4.6,4.7,10.3,8.5,16.9,11.2%20c6.6,2.8,13.9,4.2,21.7,4.2c22.6,0,36.6-10.3,45-19.9l-19.2-14.6c-4,4.8-13.6,11.3-25.6,11.3c-7.5,0-13.7-1.7-18.4-5.2%20c-4.7-3.4-7.9-8.2-9.6-14.1l-0.3-0.9h79.5L316.5,97.7L316.5,97.7z%20M237.2,88.4c0-7.4,8.5-20.3,26.8-20.4c18.3,0,26.9,12.9,26.9,20.3%20L237.2,88.4z'/%3e%3cpath%20class='st1'%20d='M387.4,135.3c-0.5-1.2-1.2-2.2-2.1-3.1c-0.9-0.9-1.9-1.6-3.1-2.1s-2.5-0.8-3.8-0.8c-1.4,0-2.6,0.3-3.8,0.8%20c-1.2,0.5-2.2,1.2-3.1,2.1c-0.9,0.9-1.6,1.9-2.1,3.1c-0.5,1.2-0.8,2.5-0.8,3.8c0,1.4,0.3,2.6,0.8,3.8s1.2,2.2,2.1,3.1%20c0.9,0.9,1.9,1.6,3.1,2.1s2.5,0.8,3.8,0.8c1.4,0,2.6-0.3,3.8-0.8c1.2-0.5,2.2-1.2,3.1-2.1c0.9-0.9,1.6-1.9,2.1-3.1%20c0.5-1.2,0.8-2.5,0.8-3.8S387.9,136.5,387.4,135.3z%20M385.8,142.3c-0.4,1-1,1.9-1.7,2.6c-0.7,0.7-1.6,1.3-2.6,1.7s-2,0.6-3.2,0.6%20c-1.1,0-2.2-0.2-3.2-0.6c-1-0.4-1.9-1-2.6-1.7s-1.3-1.6-1.7-2.6c-0.4-1-0.6-2-0.6-3.2c0-1.1,0.2-2.2,0.6-3.2s1-1.9,1.7-2.6%20c0.7-0.7,1.6-1.3,2.6-1.7s2-0.6,3.2-0.6c1.1,0,2.2,0.2,3.2,0.6c1,0.4,1.9,1,2.6,1.7s1.3,1.6,1.7,2.6c0.4,1,0.6,2,0.6,3.2%20C386.5,140.3,386.2,141.3,385.8,142.3z%20M380.2,139.9c0.8-0.1,1.4-0.4,1.9-0.9s0.8-1.2,0.8-2.2c0-1.1-0.3-1.9-1-2.5%20c-0.6-0.6-1.7-0.9-3-0.9h-4.4v11.3h2.1v-4.6h1.5l2.8,4.6h2.2L380.2,139.9z%20M379.1,138.3c-0.3,0-0.6,0-1,0h-1.5v-3.2h1.5%20c0.3,0,0.6,0,1,0c0.3,0,0.6,0.1,0.9,0.2c0.3,0.1,0.5,0.3,0.6,0.5s0.2,0.5,0.2,0.9s-0.1,0.7-0.2,0.9c-0.2,0.2-0.4,0.4-0.6,0.5%20C379.7,138.2,379.4,138.3,379.1,138.3z'/%3e%3c/svg%3e)
3rd Paris edition · 18th worldwide · December 1-2, 2026 · CNIT Forest, Paris