MSE CSE · University of Michigan
I build autonomous AI systems that do real work: agent fleets that program, research, write, and review code around the clock.
I got into programming by shipping something real. I built and published a game on Steam while still in high school: 14,000 lines of C#, 50,000+ players across 115 countries. That set the trajectory for everything that followed.
Since then I've interned at TechSmith (C++ licensing for Snagit and Camtasia) and Charles Schwab (real-time distributed systems for fractional trading: Java, Spring Boot, Kafka). This summer I'm interning at Microsoft on Azure hyperscale networking.
Now I'm an MSE student in CSE at the University of Michigan, where I'm a Graduate Student Instructor for EECS 281 (Data Structures & Algorithms) and co-developed EECS 291, a new Python-based applied DSA course, collaborating directly with the professor on curriculum and assignments.
These days my focus is on AI systems, specifically what it looks like when agents stop being toys and start doing real work. I'm building a multi-agent orchestration platform that runs 24/7, handling programming, research, writing, and code review autonomously.
Multi-agent orchestration where specialized autonomous agents (programmer, researcher, writer, code reviewer, architect) work in parallel on real tasks. An always-on orchestrator handles task queues, model routing across Claude/GPT/Codex, circuit breaking, failure recovery, and deduplication. Runs 24/7 on a Mac Mini.
Not a chatbot. An autonomous agent that runs continuously: it manages the PAW workforce, creates and triages tasks, handles communications across Discord, sets reminders, runs scheduled health checks, and finds new work on its own. Persistent semantic memory, 30-minute heartbeats, and a real-time web dashboard (Nexus) for full visibility into the agent fleet.
Social media startup · Co-founder & lead engineer
Full-stack social app with native iOS (Swift) and a strongly-typed Python backend. Real-time chat over WebSockets, location-based events, AI integration via the OpenAI API, and a Discord plugin for multi-platform engagement. Led a 3-person dev team across multiple codebases.
Couch co-op puzzle game · Capstone
Two-player co-op where you play as living magnets. Use magnetic forces to solve physics puzzles across 10+ hand-crafted levels.
with Alex Smart, Jack Kruger, Pat Attasaranya, Chawit Vongsuly
P-05
Published on Steam · High school
Where it all started. 14,000+ lines of C# in Unity, published on Steam to 50,000+ players across 115 countries and rated Very Positive (90%). Solo developer: design, code, UI, marketing, release.
A hands-on course on building autonomous AI agents: agent architectures, tool use, memory systems, multi-agent orchestration, and deployment.
Graduate Student Instructor
Teaching one of Michigan's core CS courses: discussion sections, office hours, and project support.
Co-developer
Co-developed a new Python-based applied DSA course, working directly with the professor on curriculum and assignments.