avatar

Francesco Capuano

ML Researcher
LeRobot, Hugging Face
fc.francescocapuano@gmail.com

News ๐Ÿ—ž๏ธ

[June 2025] We released a sub-500M VLA trained exclusively on open data. Check out the preprint!

[May 2025] Our paper on RL for High-Power Lasers control has been accepted to the main track at RLC'25. See you in Edmonton ๐Ÿ‡จ๐Ÿ‡ฆ!

[April 2025] I am joining the robotics team at Hugging Face to do research with Remi Cadene at LeRobot.

[March 2025] I am heading to the ELLIS Winter School on Foundation Models in Amsterdam. It's my first time in Amsterdam! Check the blogpost!

[January 2025] I joined Mistral AI working with Sophia Yang on open-source AI. Part-time while at MVA.

[December 2024] Alongside Gabriele and Caleb, we won our 3rd hackathon in a row! We closed our hackathon season with a banger, check the blogpost!

[September 2024] I started the MVA program (applied maths, Deep Learning) at ENS, in Paris ๐Ÿ‡ซ๐Ÿ‡ท.

[August 2024] I am heading to the 20th DLRL Summer School in Toronto. It's my first time in Canada ๐Ÿ‡จ๐Ÿ‡ฆ!

[April 2024] I am leaving Bain and moving to Paris! I will be joining Alan working with the CTO, Charles Gorintin.

[October 2023] After 2+ years, I got my blue belt in BJJ! ๐Ÿฅ‹๐Ÿ”ต Check the blogpost!

[July 2023] I have graduated from Politecnico di Torino, with a MSc thesis in Deep RL applied to Hardware-Aware Neural Architecture Search. Check the preprint!

[May 2023] I joined Bain & Company to work in their startup incubator as a ML Engineer ๐ŸŒ.

[April 2023] Our paper on RL for Laser Systems has been accepted to SPIE 2023. I am heading back to Prague to present our work! Edit: we won the Best Student Paper Award!

[October 2022] My first paper is out! Published in the proceedings of JaCoW 2022. Check the preprint!

[July 2022] I am moving to Prague, Czechia ๐Ÿ‡จ๐Ÿ‡ฟ, to join the ML team at ELI Beamlines, to work on ML methods in Laser-Wakefield Acceleration.

Home Blog

Ciao ๐Ÿ‘‹

Francesco here :) I deal with Deep Learning, with a particular interest in Deep Reinforcement Learning.

Born and raised in the suburbs of Rome, Italy ๐Ÿ‡ฎ๐Ÿ‡น I would have never thought Iโ€™d say this one day, but I like: (1) Research, (2) Reinforcement Learning, (3) Blues, (4) BJJ, and (5) travelling (usually in this order).

I am currently working at Hugging Face, making cute robots smarter alongside Remi Cadene and our amazing team at LeRobot.

Research Interests

Publications

  1. SmolVLA: A Vision-Language-Action Model for Affordable and Efficient Robotics
    Mustafa Shukor*, Dana Aubakirova*, Francesco Capuano*, ..., Matthieu Cord, Thomas Wolf, Remi Cadene* (*Core team)
    PDF Code

  2. Sim-is-more: Randomizing HW-NAS with Synthetic Devices
    Francesco Capuano*, Gabriele Tiboni, Niccolรฒ Cavagnero, Giuseppe Averta (*Corresponding author)
    PDF Code BibTex

  3. RLC'25
    Shaping Laser Pulses with Reinforcement Learning
    Francesco Capuano*, Davorin Peceli, Gabriele Tiboni (*Corresponding author)
    Reinforcement Learning Conference, 2025
    PDF Code BibTex

  4. FiO'24
    High-Power Laser Pulse Shape Optimization with Hybrid Stochastic Optimization Algorithms
    Ishraq Md Anjum *, Davorin Peceli, Francesco Capuano, Bedล™ic Rus (*Corresponding author)
    Frontier in Optics
    BibTex

  5. SPIE'23
    TempoRL: laser pulse temporal shape optimization with deep reinforcement learning
    Francesco Capuano*, Davorin Peceli, Gabriele Tiboni, Bedล™ic Rus (*Corresponding author)
    SPIE Optics and Optoelectronics, 2023
    PDF Code BibTex Best Student Paper Award

  6. PCaPAC'22
    Laser pulse duration optimization with numerical methods
    Francesco Capuano*, Davorin Peceli, Gabriele Tiboni, Alexandr ล paฤek, Bedล™ic Rus (*Corresponding author)
    Personal Computers and Particle Accelerator Controls, 2022
    PDF Code BibTex