CV

Basics

Name William Huey
Email willhu003@gmail.com
Url https://www.willhuey.com
Summary Computer Science student at Cornell University interested in safe interactive robotics

Work

  • 2024.09 - Present
    Cornell People and Robots Teaching and Learning Lab
    Research Associate
    Imitation learning from videos under Prof. Sanjiban Choudhury
    • Discovered a novel approach for learning difficult robotic control tasks from a video
    • Achieved >4x improvement above existing baselines on 16 manipulation and humanoid control tasks in simulation
    • Presented at a CoRL 2024 workshop and submitted a first-author publication to a top ML conference (in review)
  • 2024.06 - 2024.09
    Amazon Web Services
    Software Engineering Intern
    Cloud infrastructure for packaging and builder tools
    • Deployed an AWS microservice to optimize software package storage, projected to save **$50M+** per year
    • Developed the application in Java and Typescript using ECS Fargate, DynamoDB, SQS, S3 event notifications, Cloudwatch metrics and alarms, and internal CI/CD pipelines
    • Collaborated with developers across multiple teams to identify key design choices
    • Wrote extensive unit, integration, and load tests that reflect actual customer request patterns
  • 2023.06 - 2023.08
    ETH Zurich Robotic Systems Lab
    Research Intern
    Mobile robot traversability mapping under Prof. Marco Hutter and Prof. Donald Greenberg
    • Built a ROS package for mobile robot traversability mapping, resulting in a successful deployment on an ANYmal quadruped
    • Proposed a novel method to predict environment traversability given language constraints using VLMs
    • Distilled large VLMs into smaller visual networks capable of real-time traversability segmentation
    • Presented work at the Rawlings Cornell Research Scholarship Grand Slam, winning first prize of $500
  • 2022.05 - 2022.08
    NASA
    Software Engineering Intern
    Combinatorial methods for flight control software verification
    • Developed a Python application to generate minimal test suites achieving high combinatorial coverage in state-based control systems
    • Identified unit tests improving combinatorial coverage of Starship HLS flight control software
    • Collaborated closely with NASA engineers and NIST researchers to evaluate integration testing procedures

Education

Awards

Publications

Skills

Programming Languages
Python
Java
Typescript
Technologies
Pytorch
ROS
Git
Docker
AWS
SQL
Distributed Training
CUDA