Mert Ünsal

I am an ESOP scholar at ETH Zurich and a researcher in Project Numina where we received a 3 million euro grant from XTX Markets to build open source math intelligence.

I am interested in building safe superintelligence. I like search, autoformalization, and reinforcement learning using LLMs. Currently, I am doing some applied research on RL for theorem proving and autoformalization in Lean 4.

Update: I will be working at DRW London as a Quantitative Research Intern this summer, applying LLM reasoning to financial markets!

Previously, I did my BSc with a double major in Mathematics and Economics at Ecole Polytechnique in Paris. My previous experiences include:

• Research Assistant at SRI Lab at ETHZ where I worked AlphaIntegrator: a neurosymbolic approach for solving integrals step by step using transformers.
• Research Internship at Huawei Technologies Paris Research Center where I worked on flexible neural network training algorithms and presented the work in NeurIPS 2023.
• Quantitative Trading Intern at SIG where I worked on option pricing with stochastic volatility models.
• Member of the National Physics Olympiads team of Turkey in 2019, 2020, and 2021 where I travelled to Vietnam, Australia, Israel, Russia and won a few medals.

In my free time I play blitz or bullet chess, do any type of sports, and write blog posts. I love non-trivial conversations.

Feel free to reach out to me for any reason. I am usually quite responsive.

Email  /  Github  /  LinkedIn

Template from this website.