stephen

hung.

about.
read.cv ↗

i build AI agents i’d actually want to use. voice, vision, onchain.

Stephen Hung at the OpenAI hackathon

Berkeley EECS ’28. Built opal, darwin, lapis, yolodex, vima, iris, flow, bip, clarifai, tenzin. Won 12 hackathons doing it. $250k at pump.fun. OpenAI Codex top 5, demoed to Sam Altman. B@BHacks 1st. Incoming AI/ML intern at Walmart Global Tech this summer.

services.
  • Agents
  • Interfaces
  • Infra
  • Automation
  • Systems
Walmart
·
OpenAI
·
NVIDIA
·
Anthropic
·
Gemini
·
Y Combinator
·
pump.fun·
Cal Hacks·
MLH
·
Solana
·
Sui·
ElevenLabs
·
LiveKit
·
XRPL
·
Berkeley·
Walmart
·
OpenAI
·
NVIDIA
·
Anthropic
·
Gemini
·
Y Combinator
·
pump.fun·
Cal Hacks·
MLH
·
Solana
·
Sui·
ElevenLabs
·
LiveKit
·
XRPL
·
Berkeley·
Stephen at Y Combinator
Stephen at Yosemite
stack.
ReactNext.jsTypeScriptTailwindGSAPThree.jsFastAPIBunConvexExpressClaude APIGeminiOpenAI CodexLiveKitElevenLabsYOLOSAMCLIPmanimCOLMAPSui MoveXRPLSolanax402VercelRailwayDocker
experience.
Walmart Global Tech
ai/ml engineer intern
incoming ai/ml intern on the international digital experience team in sunnyvale. computer vision work covering human gesture recognition from realtime video feeds and person re-identification across multi-camera in-store footage.
summer 2026, incoming
Enactus Berkeley
tech lead
tech lead for social impact ventures across multiple chapter teams.
since spring 2026
Cal Blueprint (Replate Food Rescue)
software developer
building react native + expo app replacing replate's pen and paper donation logging across a network rescuing 1m+ lbs of surplus food annually for 80+ recipient nonprofits. architected token based auth extending their ruby on rails backend.
since sept 2025
ClearPath Medical
software engineering consultant
built a pfmea automation tool with fastapi + llms that converts work-instruction pdfs into risk assessments. cut analysis time for fda-compliant medical device qc by 85%. websocket job tracking, 4-phase validation pipeline, local sqlite to keep phi off the network.
fall 2025
Theta Tau
professional development chair
professional engineering fraternity at uc berkeley. ran pro dev programming for the active body.
fall 2025
OptiGenix
software engineering intern
trained a generative ai model on google vertex ai to 92.3% extraction accuracy on 60 unstructured blood test pdfs. built the gcp pipeline behind 7,500+ monthly pdf uploads. iam scoping, encryption at rest, audit trails so the hospital side passed compliance review.
summer 2025
faqs.

EECS at UC Berkeley, class of '28. Started shipping in middle school with a JavaScript Mother's Day card, then VEX robotics captain (PID + odometry for autonomous routines), then FBLA nationals (mobile + computer programming), now hackathons. Valedictorian out of Ayala HS, GPA 4.69. Currently dev at Blueprint (tech for nonprofits) and previously SWE intern at OptiGenix (Skydeck Pad-13 alumni).

Honestly, whatever's pulling me. Random corporate and collegiate hackathons every few weeks, side projects with friends that start in a discord call at 2am, design work for people who vibe with my taste, and a rotating stack of personal builds I'll ship when they're good. Always one project shipping, one being rebuilt, one fresh, usually too many at once. If something's interesting enough to break my sleep schedule for, I'm probably already on it.

Yes. Booked for Walmart Global Tech this summer, but open for offseason internships (spring/fall '26) right now, plus summer 2027 and post-grad full-time (May 2028). Most interested in agentic systems, generalist SWE, frontend/design, honestly anywhere I can ship something real. Email me with the role and I'll reply within 48h.

If it's a hackathon, pick something that demos in 90 seconds and could keep going for a year. If it's a personal project, pick the part of the stack I haven't shipped before. Either way, the rule is novel tech + a constraint that forces taste. Lapis came from "can crypto conditions enforce vesting without lawyers," Opal came from "can an agent actually queue up with you," Yolodex came from "can you turn a YouTube clip into a trained YOLO model overnight."

Bun > npm. Vite > CRA. TypeScript strict. FastAPI for the backend, Next.js for the frontend, GSAP for motion. AI work in Python or Bun depending on which one's faster to ship. I optimize for how fast I can rebuild something, not how clever the architecture looks.