Skip to content

About

Software Engineer / Computer Science Student — based in Ann Arbor, MI.

I'm a computer science student at the University of Michigan. This summer I interned at CodeHS on Bool, an AI app builder: a prompt becomes a hosted web app. Most of the six weeks went into the parts that fail silently — a credit meter that wasn't measuring cost, a deploy path agents couldn't finish, and login links that mail scanners burned before a person could click them.

I'm also the AI lead on UM::Autonomy, where I run a 30-person team across computer vision, navigation and controls, and task planning. A lot of what I know about software came from watching a perception pipeline fail on the water and having to explain why.

I like the parts of the stack most people skip: memory layouts, message boundaries, the reason a thing is slow. That shows up in a 2D game engine I wrote in C++ that resolves collisions between 100,000+ entities, and in a telemetry layer built on Zenoh and Protobuf that keeps 15+ distributed ROS2 nodes talking to shore.

I work end to end when the project calls for it — the photo-sharing site I maintain is React and Next.js on top of a custom OAuth flow I designed against Firebase and Adobe's APIs. I care about software that's still understandable a year after it ships.

stack

Languages

  • C
  • C++
  • Python
  • Java
  • TypeScript
  • JavaScript
  • Lua

Robotics & Systems

  • ROS2
  • Zenoh
  • Protobuf
  • Gazebo
  • RViz
  • SDL2
  • Box2D

Web

  • React
  • Next.js
  • Tailwind CSS
  • Firebase
  • OAuth

Tools & Data

  • Git
  • Docker
  • Linux
  • GitHub Actions
  • NumPy
  • Pandas

Certifications

  • CompTIA A+
  • Network+
  • Security+

Timeline

  1. June 2026 — August 2026

    Software Engineering Intern

    CodeHS

    • Turned a prototype in-app AI meter into a priced product: one credit pool, real provider cost, and a daily ceiling so one public app cannot empty a workspace.
    • Rebuilt Bool Drop so an agent or a browser can upload a site and get a live URL, and made the hard plan limits agents used to retry forever actually say why they failed.
    • Fixed the silent failures — a site-wide outage from a pool that never drained, an editor that never left the loading skeleton, and login links that mail scanners burned before a person could click.
  2. January 2025 — Present

    Computer Consultant II

    University of Michigan

    • Debug the hardware–software interface across 30+ campus facilities, diagnosing RS232 serial communication and network topology failures.
    • Run real-time incident response over Slack, tightening how resources get allocated during outages.
  3. October 2024 — Present

    AI Lead

    UM::Autonomy

    • Lead a 30-person AI team across three subteams — computer vision, navigation and controls, and task planning — with weekly design reviews and code walkthroughs on every pipeline.
    • Directed integration of deep-learning object detection and LiDAR point cloud classification in ROS2, Python and Docker, improving perception reliability for competition tasks.
    • Designed a low-latency telemetry and command architecture on Zenoh and Protobuf for reliable bidirectional traffic across 15+ distributed ROS2 nodes.
  4. Expected December 2026

    B.S.E. Computer Science

    University of Michigan, Ann Arbor

    • Coursework in data structures, computer organization, practical data science, linear algebra and cybersecurity.
    • CompTIA A+, Network+ and Security+ certified.