Machine Learning for Computer Vision using Counterfactual World Models (CWMs), a vision transformer-based foundation model. Working on optimizing counterfactuals for extracting optical flow using CWMs.

Designed signal extraction tools for fixed-income trading.

Studied random quantum circuits to bound the depth at which they form approximate unitary designs. Combined theoretical calculations with numerical computations in Python to diagonalize local quantum Hamiltonians in order to compute their spectral gaps. Sought to improve current depth bounds and extend to broader classes of circuits.

Modeled potential gravitational wave signals from high-redshift supermassive black hole binaries. Wrote Python notebook to analyze astronomical data and estimate supermassive black hole masses, then calculate potential strain amplitudes, and finally compare between low- and high-redshift samples. Learned SQL to query astronomical databases. Publication in progress.

Studied diHiggs VBF and ggF simulations through the bbγγ decay mode with the SLAC ATLAS group. Wrote C++ code using ROOT library to analyze simulated ATLAS data. Used Python to compare results across different simulations and determine upper limits on a fundamental physical parameter (c2V) of elementary particle interactions.