Hello there.

I am a Research Scientist at MiroMind, working on large language models, long horizon reasoning, and agentic systems.

Over the past year, I led the effort on agentic preference optimization for the MiroThinker series. Welcome to try it out here 🪞✨.

Previously, I was an NLP Quant Research Scientist at GIC, and co-founded a legal AI startup, where I helped define the core research problems and built systems for contract understanding and reasoning.

I am particularly interested in developing an AI assistant’s sense of taste, how it learns to recognize and internalize what "good" looks like. I don't quite know where to start yet, but I find that part of the problem especially fun.