The short version of my working life. The rest of the site goes deeper on teaching, ICPC, and research.
I build systems that must remain reliable under real-world pressure—and the evaluation and feedback loops that make them improve.
Fifteen years across Google, Two Sigma, and startups; currently Chief Architect at Probook. My work spans production AI architecture, empirical evaluation, mathematical research, and technical organizations. I also teach at Columbia and direct the ICPC Foundation’s internship program.
Worked on execution systems for systematic trading strategies — measuring and improving how orders actually get filled.
Applied AI research that shipped into Google products, and core infrastructure for large-scale ML research. Tech lead for Google Kick Start, which grew from a regional contest into a competition with 110K+ participants worldwide.
A reusable framework for making explicit, reviewable choices across quality, cost, risk, delivery time, and reversibility without publishing employer-specific measurements.
Read the framework →Combined exact hardness, approximation, structural analysis, counterexamples, and rational-arithmetic computation in an active research program with Josh Alman.
Research overview →Built validation discipline around high-stakes problem packages: independent solutions, adversarial submissions, generated data, automated checks, and human sign-off.
Engineering overview →I run the Foundation’s internship program — 136 completed internships and 14 interns active as of August 14, 2026, across projects like the ICPC Archive, mobile app, and online training platform. I also chair the Competitive Learning Institute Symposium at the World Finals and co-chair an open-source curriculum covering 50+ algorithmic topics (50K+ downloads).
Fourteen course offerings through Fall 2026 in competitive programming and interview preparation, reaching 500+ students — evaluation results published in full, rough semesters included. 100+ individual mentees, with placements verified and listed.
Launched the inaugural issue (Vol. 1, No. 1, May 2025), featuring a co-authored review article, “Programming Contests, Competitive Programming Curricula, and how Everybody Benefits.”
Research on network effects, information cascades, and game theory in complex systems.
Focus on machine learning, computer vision, and scalable AI systems.
The Journal of Competitive Learning, 1(1) · 2025
Proceedings of the ACM on Programming Languages, 8 (OOPSLA1), 810–832 · 2024
ACM SIGMETRICS Performance Evaluation Review, 44(3), 7–14 · 2017
International Journal of Computer Vision, 107(1), 20–39 · 2014