experience

Full resume available upon request

Professional

Software Engineering Intern @ Meta

June 2024 - August 2024

  • Created an operating system agnostic network topology synchronization pipeline from scratch in C++ and Thrift to support 1.2M+ devices in data centers
  • Deployed in production, reduced onboarding time for New Product Introduction by 3x, improved latency by 31%

Software Engineering Intern @ Wells Fargo

June - August 2023 | Charlotte, NC

  • Created a programmatic benchmarking framework for LLMs in Python using PyTorch, Tensorboard, and Weights & Biases, focusing on evaluating memory footprint for the certification process of models in 45+ GenAI use cases
  • Fine-tuned FLAN-T5 via PEFT, achieving a ~97% smaller model with ~98% performance of full-tuned model

Quantitative Developer @ Stealth

June - November 2023

  • Algorithmic trading in Python via Polygon and Alpaca market APIs

Software Engineering Intern @ Golden Eagle Partners

June - August 2022 | Denver, CO

  • Implemented a dynamic lookback-based breakout and trailing stop loss trading strategy via an algorithmic trading bot in Python that resulted in 12% returns off a simulated portfolio of $100,000 diversified across 5 industries

Summer Equity Analyst @ Somar Capital Management

May - July 2022 | New York, NY

  • Analyzed investment opportunities within the European and American aerospace and defense markets by constructing DCF models

Research

Machine Learning Researcher @ Harvard Data to Actionable Knowledge Lab

September 2023 - Present | Cambridge, MA

  • Exploring the applications of inverse reinforcement learning in the medical setting in a 3-person team under the supervision of Prof. Finale Doshi-Velez

Research Assistant @ Harvard Programming Langauges Lab

September 2023 - February 2024 | Cambridge, MA

  • Investigated reasoning capabilities of large language models under the supervision of Prof. Nada Amin
  • Collaborated on a novel Monte Carlo Tree Search system in Python to guide LLMs in generating verified programs in Rust, Scala, Coq, Lean, and Dafny
  • Investigated using probabilistic belief graphs, information identification, and logical constraints to enhance LLM reasoning among various BIG-bench benchmarks

Extracirricular