I’m a computer science PhD student at CU Boulder, advised by Ramin Ayanzadeh. I’m interested in what it takes to run useful quantum computations on real hardware, and how AI can help improve quantum computers. I enjoy algorithmic puzzles, building software, and making computers do what I want.
For my undergraduate thesis, we compared quantum and classical support vector machines using Dhaka Stock Exchange data. We explored how performance changed with dataset size, feature count, and quantum kernel design.
In this undergraduate research project, we compared quantum and classical kernels for leukemia classification using gene expression data. We examined their performance across different data settings and estimated the quantum resources needed for larger experiments.
We tested whether giving LLM agents access to Qiskit documentation could help them generate better code. Our system split the work between planning, coding, and execution agents. On Qiskit HumanEval, it improved Pass@1 by up to 15.9 percentage points over single-pass generation on the mid-sized open-weight models we tested.
At Re:tune (Nov 2023 to Jun 2026), I worked on an AI coding product. I built its browser editor and the systems for versioning files, saving project snapshots, and coordinating AI agents. I also helped build SierraRomeo, an aviation research assistant that answers questions using more than 60,000 FAA documents and links to its sources.
I also worked on 10xlaunch, a tool for identifying website visitors. It ranked #1 Product of the Day and among the top five products of the month on Product Hunt in August 2024.
