Building AI agents and AI-driven workflows to accelerate product velocity

Published ConMe (https://arxiv.org/abs/2406.08164), which was recently accepted at NeurIPS 2024 Datasets and Benchmarks Track

Taught a recitation section of 20+ students twice every week and hosted weekly Office Hours to assist students, for the course 18.C06 Linear Algebra and Optimization; recorded a midterm review video which totaled 140 times and 2398 minutes viewed

VCMM (Virtu Client Market Making) desk

Utilized Large Language Models to automate the sensemaking pipeline of conversations for Cortico (https://cortico.ai/), particularly prompting models from the GPT-3.5 family to perform topic tagging