resume

OpenAI

Incoming Member of Technical Staff Intern

• Incoming Summer 2026.

Stanford Artificial Intelligence Lab

Research Assistant — Professor Yejin Choi

VERL, vLLM, Slurm

• Built a multi-agent RL pipeline using VERL with Slurm-managed training on NVIDIA H100s; trained shared-parameter LMs using dual reward functions for quality and diversity, improving Shannon Evenness Index by over 15%.

Stanford School of Engineering (CS198)

Teaching Assistant

• Lead sections of 15 students for Stanford’s DSA class. Grade assignments and debug Python & C++ code.

MultiOn - AGI, Inc

AI Engineer Intern

LangChain, LangGraph, Chrome DevTools

• Designed a benchmark for evaluating WebAgents and prototyped a memory-augmented browser agent using LangChain, LangGraph, and OpenAI APIs; fine-tuned the agent with RLHF via GRPO, over 10% reduction in average latency.

Stanford Carta

Backend Engineer

Docker, C#, SQL, .NET Core

• Reduced p95 API latency 30% across 4 .NET Core services powering the web app (~2k MAU) by adding read-through Redis caching and tuning SQL connection pools
• Built a Dockerized SQL ingestion pipeline handling ~100k records/week (bulk upserts, idempotency, retry/backoff) with 99.9% success and sub-second median latency

Jane Street Capital

Academy of Math and Programming

• Applied optimization theory and time series data to develop trading software using Python and statistical models.
• Analyzed problems in discrete mathematics, probability, and theoretical computer science, designing stochastic models.
• Won 1st place in the Mystery Planet Puzzle Hunt and 2nd in Trading Competition; awarded $5,000 merit scholarship.