Alex L. Zhang

Hi 👋! I broadly work on areas where language models are underutilized or inefficient. Read my latest work on Recursive Language Models (RLMs). I am part of the core team running the GPU MODE leaderboard, where we are currently hosting the NVFP4 Blackwell competition with NVIDIA, and recently hosted two $100k competitions with AMD and a model optimization competition with Jane Street.

I am a PhD student at MIT CSAIL advised by Omar Khattab and Tim Kraska. I spent the summer before at Sakana AI in Tokyo, Japan. I spent a year before that as a researcher at VantAI in NYC working on AI-based drug discovery.

Before that, I graduated as the top student of the Princeton CS department, where I was blessed with amazing mentors: Professor Karthik Narasimhan, Dr. Khanh Nguyen, Dr. Ofir Press, and Professor Kai Li. During that time, I was doing an assortment of AI/ML research while interning at Snapchat, Apple , and Claryo. Before that, I used to make and sell PC games, one of which was mildly successful (~100k+ players). [unfinished example]

research highlights:

• ML Systems. Recursive Language Models (RLMs) (2025), Project Popcorn 🍿 (2025), Triton kernels for OSS AlphaFold3 (1k+ ⭐'s) (2024)
• Language Model Benchmarks. VideoGameBench (2025), KernelBench (ICML, 2025), SWE-bench Multimodal (ICLR, 2025)
• Models. Neo-1 (2025), KernelLLM-8B (2025), Language-guided World Models (SpLU-RoboNLP @ ACL, 2024)

Feel free to reach out to talk about anything through my email at [x]@mit.edu where [x]=altzhang. Finally, I’m also fairly active in and help moderate the GPU MODE community.