HomeCompaniesThesis Labs

Thesis Labs

Thesis is an applied research lab accelerating human-in-the-loop AI…

What if we could discover the next Transformer, invent the next AlphaFold, 10x faster? We are building systems that collaborate with humans to invent, train, and refine machine learning models. These agents operate in continuous, self-improving loops, guided by human creativity and intuition. Together, they accelerate the entire lifecycle of AI R&D, from idea to implementation.
Active Founders
Luigi Charles
Luigi Charles
Co-Founder
Engineering @Thesis. Betting the farm that AI can discover new science. Zero hedge. Thesis is inevitable. Second time founder. Previously co-founded Sphere. Scaled it's fintech rails from zero to 1B+ in annual global payments volume before everyone realized stablecoins were useful. SWE and equity research analyst in a past life. If you do cannot build it from scratch, you do not understand it.
Sergio Charles
Sergio Charles
Founder
I'm deeply interested in accelerating AI for scientific discovery. Before Thesis, I was an AI Resident at Google X and did an applied deep learning research internship at Nvidia. With an M.S. in Statistics and a B.S. in Math and Computer Science from Stanford, my background spans pure math, reinforcement learning, and natural language processing. I’ve contributed to research and projects ranging from climate change modeling in Andrew Ng's ML group to autonomous RL in Chelsea Finn's group.
Gil Kornberg
Gil Kornberg
Founder
Before Thesis, I operated in small deeptech startups and did AI research. My past research experience includes stints at the Stanford AI Lab using AI to tackle problems related to climate change, in Nobel Laureate Michael Levitt's Lab at the School of Medicine working on neural representations of proteins, and at TAU where I investigated the application of number theory to neurophysics. I obtained my BS (Hons.) and MSCS at Stanford.
Thesis Labs
Founded:2025
Batch:Fall 2025
Status:
Active
Location:San Francisco
Primary Partner:Tom Blomfield