Student at Cary Academy. interested in ece and physics.
MTS @ NiLabs, where we build testing tools for the software that runs AI models.
Github activity:
Projects
-
cf-recommend
Tells you what Codeforces problem to do next.
-
papermarket
Prediction market, fake money. Try it.
-
conform
Checks whether inference engines actually agree. More.
-
ladder
Small models fine-tuned to do competitive programming.
-
Teammind
Team memory for Claude Code.
-
openvim
Neovim, but with agents.
-
Claude Code Wrapped
Spotify Wrapped, but for your terminal.
Rest is on GitHub.
Contributing/maintaining:
-
OpenShell
Safe, private runtime for autonomous agents.
-
vllm-project/vime
Post-training framework