PeopleofSuperintelligence

Researcher

Gabriel Synnaeve

Gabriel Synnaeve is an AI researcher whose work spans game AI, speech recognition, object detection and code models. He studied computer science at Grenoble INP’s ENSIMAG from 2006 to 2009. In 2012, he pursued Bayesian game-AI research with Pierre Bessière and coauthored a StarCraft dataset paper. He coauthored Wav2Letter, DETR, Code Llama and Code World Model, and published a 2026 study on reinforcement learning for code optimization. As of October 2026, he is co-founder and co-CEO of Aire, an AI startup aiming to build an AI Scientist.

Now at
Aire
Badges

Timeline

  1. 2006–2009

    Computer science at ENSIMAG

    He studies computer science at Grenoble INP’s ENSIMAG from 2006 to 2009.

  2. 2012

    Bayesian game AI research

    He pursues Bayesian-programming research for multiplayer games as a PhD student in INRIA Rhône-Alpes’ E-Motion team, under Pierre Bessière at Collège de France’s LPPA. In 2012, he and Bessière publish a StarCraft dataset and army-clustering paper at an AIIDE workshop.

  3. 2016

    Wav2Letter speech recognition

    In 2016, he coauthors Wav2Letter, an end-to-end speech-recognition system that pairs a convolutional acoustic model with graph decoding.

  4. 2020

    DETR object detection

    In 2020, he coauthors DETR, which frames object detection as direct set prediction and uses a transformer encoder-decoder with bipartite matching.

  5. 2023

    Code Llama for code

    In August 2023, he coauthors Code Llama, a family of large language models for code based on Llama 2.

  6. 2025

    Code World Model

    In 2025, he coauthors CWM, a 32-billion-parameter open-weights language model for research on code generation with world models.

  7. July 2026

    Reinforcement learning for code

    In July 2026, he coauthors a study that makes code execution time a learnable reward through reinforcement learning, with DMC-Optim tests and a calibrated sandbox.

  8. As of October 2026

    Co-CEO of Aire

    As of October 2026, he is co-founder and co-CEO of Aire, an AI startup aiming to develop an AI Scientist. The company automates model building from data through deployment.

Career

  1. NowAireCo-founder and co-CEO

Education

  1. 2009Institut polytechnique de Grenoble
  2. InriaPhD student

From the biography.

  1. People shaping AI work at a median of 3 organizations over a career: 3 for university starters and institute and lab starters, versus 2 for company startersIncludes Inria. Across 6,853 people shaping AI with a dated work role, we estimate a median of 3 parent organizations over a career (mean 3.6), comparing role groups, starting sectors and starting decades from the 1950s to 2020–2026.
  2. The company share of AI researchers’ top papers was 31% in 2020–2024 and 8% in 1970–1974Includes Inria. We estimate where the people shaping AI worked when their most-cited papers appeared: the company share was 31% in 2020–2024 and 8% in 1970–1974, and Google led all organizations in 2020–2024.
  3. Google has the highest combined h-index among AI organizationsIncludes Inria. We estimate that Google holds the highest combined h-index at 8,016, while universities account for 73% of the deduplicated total across AI organizations.
  4. Companies supply the most prior-work hires at 7 of the 9 frontier AI labsIncludes Inria. Our estimates of the organizations where 550 people worked directly before joining 9 frontier AI labs across 620 career transitions, comparing the main feeder organizations per lab.

Sources and credits

Photo: We have no published photo of Gabriel Synnaeve yet. Send us a free-licensed photo.

  1. Gabriel Synnaeve's homepage emotion.inrialpes.fr. Checked 21 October 2026.
  2. A Dataset for StarCraft AI and an Example of Armies Clustering | Proceedings of the AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment ojs.aaai.org. Checked 21 October 2026.
  3. Wav2Letter: an End-to-End ConvNet-based Speech Recognition System arxiv.org. Checked 21 October 2026.
  4. End-to-End Object Detection with Transformers arxiv.org. Checked 21 October 2026.
  5. Code Llama: Open Foundation Models for Code arxiv.org. Checked 21 October 2026.
  6. CWM: An Open-Weights LLM for Research on Code Generation with World Models arxiv.org. Checked 21 October 2026.
  7. Reinforcement Learning for Code Optimization arxiv.org. Checked 21 October 2026.
  8. Tra guida autonoma e ingegneria degli atomi l’IA diventa “fisica” italian.tech. Checked 21 October 2026.