Skip to content
View Joh221105's full-sized avatar
🎯
Open to new-grad SWE roles (2027)
🎯
Open to new-grad SWE roles (2027)

Block or report Joh221105

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
Joh221105/README.md

John Huang

ML systems and full-stack engineer. M.S. in Technology Innovation at the University of Washington GIX (graduating March 2027), with a B.S. in Computer Science and a background in biochemistry.

Most of my recent work sits where AI meets real constraints: exact solvers where heuristics time out, pipelines that have to be cheap enough to run every day, and systems that keep a person in the loop on every AI decision.

Portfolio · LinkedIn · johnhuang105@gmail.com


Featured: ENGINExLaunch (private repo)

Capstone operations platform for two University of Washington programs, GIX Launch and ENGINE: team formation, mentor matching, progress tracking, and presentation feedback. One app built for 150+ students and 140+ mentors, on 8 backend services with 938 tests behind a 12-job CI pipeline.

  • Team formation on an exact solver. OR-Tools CP-SAT under lexicographic priorities reaches a provably best team assignment in under 10 seconds, where the old heuristic timed out at 60.
  • Multi-persona critique panel. One pass builds a digest of a team's deck, reviewer personas critique it, and an orchestrator synthesizes their critiques into one set of feedback.
  • Monorepo merge. Reunited two codebases that had diverged for months under one rule: share mechanism, never policy. CI checks every service against typed agent contracts, and the merge removed 46K lines of dead code with no loss of test coverage.

Python FastAPI OR-Tools Supabase ChromaDB Next.js TypeScript GitHub Actions

Read the write-up

Professional work (private repos)

Secondsite · Software Engineering Intern, Computer Vision Automated pipeline turning crane-mounted camera footage into survey-aligned aerial maps, migrated from an 8 GB Jetson to AWS. Cut cost per site map by 75% ($11.80 to $2.90) across 11 coordinated Docker services, and delivered the first survey-aligned map at 1 cm per pixel. Read the write-up

T-Mobile · Thoth Project Knowledge-capture RAG system that records subject-matter experts' knowledge through guided interviews and cites its source on every answer. Placed first among competing teams on instructor black-box benchmarks. Routing short decisions to a cheaper model cut p95 latency from 19.3 s to 6.9 s at 59% lower cost. Read the write-up

Projects

VESPR · Bat-Inspired Perching Drone (UW GIX capstone, in progress) Onboard vision that lets a quadcopter find a branch, decide whether it can grab it, and line up to perch on it, so it can cut its motors and keep watching. The target use is wildfire overwatch. Earlier perching research located the branch with external motion capture; VESPR does that sensing onboard. As perception lead, I own everything from camera pixels to a 3D branch axis and diameter that the flight controller can act on, running on a Jetson Orin Nano within a set latency budget. Python OpenCV pyrealsense2 Jetson Orin Nano RealSense D435i pytest DVC Read the write-up

Fathom (private repo, in development) iOS focus app built around fishing: cast a line, put the phone down, and the longer you stay away, the deeper the line sinks. A native Swift module tells locking the phone (allowed) apart from leaving the app (starts a 30-second fuse) without Screen Time permissions. In on-device testing, it misread none of the 22 leave-the-app runs as a lock. The server credits only the smallest of the claimed, committed, and server-metered minutes, so changing the phone's clock earns nothing. Offline-first, with 1,281 unit tests and row-level security on all 18 tables. It's a solo build using a Claude Code subagent workflow, where I own the spec, the architecture, on-device testing, and review. TypeScript React Native Expo Skia Swift Supabase PostgreSQL Read the write-up

InboxIQ AI email triage: Gmail sync, focus-based sorting, and deadline and pickup-code extraction, with a background worker that falls back to a second model provider when the first returns an unusable answer. Row-level security on every table, 325 tests, and an accuracy gate of 85% or better on hand-labelled emails. Next.js TypeScript FastAPI Supabase Gmail API Claude API Read the write-up

Toolbox

Python · TypeScript · SQL · FastAPI · Next.js / React · PostgreSQL · Docker · AWS (EC2, EBS) · GitHub Actions · OR-Tools · OpenCV · PyTorch

Popular repositories Loading

  1. InboxIQ InboxIQ Public

    AI triage for a busy inbox: sorts Gmail by your current focus and pulls out deadlines and pickup codes. Next.js, FastAPI, Supabase.

    Python

  2. Joh221105 Joh221105 Public