Sahil Chowdhury
I’m a Turing Scholar at UT Austin studying mathematics and computer science. This summer I’ll be at Optiver as a quantitative developer, working on the systems that turn research into live markets.
Before that I spent a summer at BlackEdge Capitalbuilding low-latency C++ trading infrastructure — broadcasting ML price predictions in under 30ms, shaving 15% off a hot path, and shipping the firm’s first production health monitor. Earlier internships took me to Fidelity writing disaster recovery pipelines across 5TB of data, and to MD Anderson, where I adapted recent XGBoost work to push cancer-classification accuracy from 92% to 98% on a hundred thousand cells.
What I like best are problems that have to be fast, correct, and cleverat the same time. In practice that’s meant writing a preemptive multithreaded Unix kernel, a pipelined AArch64 CPU with branch prediction, a matching engine that clears a hundred thousand orders a second, and a little language with coroutines and a hand-rolled heap. Lately I’ve been pulled toward AI/ML research — especially the parts where the systems engineering and the math stop being separable.
Outside of work I speedcube (sub-18), play chess and a fair amount of poker, and lose weekends to puzzles, rock climbing, and Minecraft.