Lucas Biéchy

Curriculum vitæ

The path so far

Physics and mathematics first, then statistics, then AI safety, with a soft spot for privacy. Now a PhD researcher at Inria, thesis expected September 2027.

Open the PDF version ↗
01

Experience

  1. 2024 - now
    Inria, PETSCRAFT team

    PhD researcher in AI safety & privacy

    Privacy in LLM agents and evaluation of what models really do: calibrated uncertainty for black-box reasoning models (via jailbreaks), detecting biased self-grading agents and biased users from behaviour, and measuring how agents handle personal data. Four first-author papers: two under review, one workshop paper at NeurIPS 2026 and one preprint.

    Palaiseau, France
    • LLMs
    • Oversight
    • Uncertainty
    • Agents
    • Privacy
  2. 20245 months
    NARLabs, National Center for High-performance Computing

    AI research intern

    Designed and implemented xGRU, inspired by xLSTM, doubling traffic forecasting accuracy and halving model parameters. Published in IEEE OJ-ITS.

    Hsinchu, Taiwan
    • Deep Learning
    • Time Series
  3. 20235 months
    Exotrail

    Data engineer intern

    Built from scratch a full data pipeline for physics experiments, from raw data to automated analysis in minutes.

    Massy, France
    • Data
    • Pipelines
    • MLOps
  4. 20222 months
    CNRS, Laboratoire de Mathématiques d'Orsay

    Research intern

    Enhanced numerical Fast Marching methods and improved statistical density estimation techniques.

    Orsay, France
    • Numerical methods
    • Statistics
02

Education

  1. 2024 - 2027
    University of Paris-Saclay

    PhD in Artificial Intelligence

    Advisors: Nicolas Anciaux, Cédric Eichler, Adrien Boiret. Thesis expected September 2027.

    Palaiseau, France
    • AI Safety
    • LLMs
    • Privacy
  2. 2022 - 2024✺ High Honors
    University of Paris-SaclayCentraleSupélec

    MSc in Mathematics (spec. in Artificial Intelligence)

    Joint program with CentraleSupélec and École Polytechnique. GPA 4.00/4.00.

    Orsay, France
    • Machine Learning
    • Statistics
    • Optimization
    • Probability
  3. 2019 - 2022✺ Honors
    University of Paris-Saclay

    Double BSc in Mathematics and Physics

    Physics track joint with École Normale Supérieure Paris-Saclay. Mathematics GPA 4.00/4.00, Physics GPA 3.33/4.00.

    Gif-sur-Yvette, France
    • Mathematics
    • Theoretical Physics
03

Papers

Four first-author papers in 2026 (two under review, one NeurIPS workshop paper, one preprint), plus one journal paper.

See all papers →
04

AI safety

05

Service & more

Visiting
Trajan House, Oxford (2026) · UC3M, Madrid (2025, 2026)
Summer school
Machine Learning Summer School, Columbia University, New York (2026)
Workshops
ELLIS/ELSA Privacy-Preserving ML, Bertinoro (2025) · APVP, Chasseneuil-du-Poitou (2025)
Teaching
Statistics & Probability TA at ENSTA Paris (2026) · Academic tutor (2023)
Mentoring
Christophe Rechdan (2026) · Adam Gassem (2025)
Talks
AI & Cybersecurity, Paris (2026) · Sovereignty & Entrepreneurship, Gif-sur-Yvette (2026)
Organising
PETSCRAFT research seminar, Inria (2025)
Outreach
Inria science mediation team (2024-2025)
Entrepreneurship
Doctoral entrepreneurship programme (2024-2027)
Grants
International Intern Scholarship, NARLabs (2024) · MixtAI Scholarship, Institut DataIA (2023 & 2024)
06

Skills

ML
  • Python
  • PyTorch
  • Hugging Face
  • scikit-learn
Software
  • TypeScript
  • React
  • PostgreSQL
  • Supabase
Languages
  • French (native)
  • English

Up for a collaboration, a talk or a coffee?

Drop me a line ↗