Available for opportunities

Hi, I'm To Yin Yu

Full-Stack EngineerApplied AI Builder

Entry-level full-stack engineer building applied AI products with TypeScript, Python, Next.js, FastAPI, WebRTC, and the OpenAI Realtime API. Recent work includes realtime voice tutors and mock-interview agents, plus a deployed analytics platform with auth, rate limiting, and Postgres.

01. about

About Me

to-yin-yu.ts
const toYinYu = {
  name: 'To Yin Yu',
  role: 'Full-Stack Engineer',
  focus: 'Applied AI · Realtime',
  edu: 'B.S. CS, UMD',
  stack: [
    'Next.js', 'FastAPI',
    'WebRTC', 'OpenAI',
  ],
  status: 'open to work',
};

I'm To Yin Yu — an entry-level full-stack engineer focused on applied AI products, with a B.S. in Computer Science from the University of Maryland, College Park.

I ship end-to-end products: Next.js and React frontends, FastAPI and Python backends, PostgreSQL data models, Redis caching and rate limiting, and realtime voice agents built on the OpenAI Realtime API over WebRTC.

I'm looking for junior software engineering roles where I can keep shipping at this pace — full-stack, with a strong applied AI lean.

B.S.
CS @ UMD
10
Deployed Projects
3
Realtime Voice AI Apps
Prod
Auth · APIs · Deploys

02. skills

Tech Stack

Technologies I use to bring ideas to life

Languages

TypeScriptPythonJavaScriptSQL

Web & Backend

Next.jsReactFastAPITailwind CSSPostgreSQLRedis

Applied AI & Realtime

OpenAI Realtime APIWebRTCElevenLabsLLM Tool UseTensorFlowKeras

Infra & Deploys

DockerGitLinux/UnixVercelRenderNeonUpstashNetlify
FamiliarRust · OCaml · Ruby · Swift · Java · C · Astro · Pandas · NumPy · sklearn · SAS

03. projects

Things I've Built

A few featured projects — see them all on the projects page.

2026featured

Throughline — Technical Learning Ecosystem

The hub for a connected set of first-principles technical learning guides — a 5-act ladder (Code → Specialize → Build → Prep → Practice) linking every guide plus SoloMock and Shipyard into one path from absolute beginner to job-ready.

Highlights

  • Designed and shipped a static Astro hub that maps the whole ecosystem as a 5-act learning ladder, routing readers from a programming on-ramp through specialization guides into building and interview practice
  • Unified every guide on one house design system (`@throughline/guide-kit`) and a single pedagogical standard — readers enter knowing nothing and leave job-ready, read top-to-bottom
  • Source-available across the ecosystem (© 2026 To Yin Yu, All Rights Reserved), with in-browser interactive code and checkpoint quizzes throughout
AstroTypeScriptDesign SystemVercel
Private repo · code available on request
2025featured

SoloYap

Realtime voice English tutor — practice spoken English with an AI partner over scenarios like ordering coffee, doctor visits, and salary negotiation, with grammar corrections after every turn.

Highlights

  • Built the realtime voice loop on OpenAI's Realtime API over WebRTC; instant grammar corrections and better-phrasing suggestions stream back after every learner turn
  • Authored 13 CEFR-graded scenarios (A2–C1) plus free-chat and custom-topic modes, with optional Mandarin/Cantonese native-language hints to lower the floor for beginners
  • Shipped as an installable PWA with manifest + apple-touch icons; layered cost protection via per-IP rate limits, session caps, and an email-allowlist request-access flow
Next.jsReact 19TypeScriptOpenAI Realtime APIWebRTCTailwind CSSPWA
Private repo · code available on request
2025featured

SoloMock

Verbal mock-interview app — talk through coding problems out loud with an AI interviewer over realtime voice, get instant feedback.

Highlights

  • Built the realtime voice loop on OpenAI's GA Realtime API over WebRTC; debounced code-editor snapshots stream into the model so the AI reacts to what you type, not just what you say
  • Authored 15 structured per-problem 'interviewer briefs' (solution tree, 4-rung Socratic hint ladder, follow-ups, edge cases) that drive interviewer behavior via the system prompt
  • Layered cost protection: per-IP rate limit, 15-min session cap, ephemeral key TTL, and a Discord-webhook extended-access request flow with manual-whitelist approval
Next.jsTypeScriptOpenAI Realtime APIWebRTCMonacoTailwind CSS
Private repo · code available on request
2024featured

all-in-one-URL

Short links, QR codes, barcodes, and analytics platform.

Highlights

  • Built FastAPI backend with PostgreSQL, Redis caching, and tiered rate limiting
  • Added JWT auth and per-resource ownership while preserving anonymous resources
  • Deployed full stack on Vercel, Render, Neon, and Upstash
PythonFastAPIPostgreSQLRedisDockerReactTypeScript
Private repo · code available on request

04. experience

Experience & Education

Independent Software Projects

2024 – Present

Self-directed · Full-stack & Applied AI

Designed, built, and deployed full-stack and applied AI products end-to-end — realtime voice tutors and mock-interview agents on the OpenAI Realtime API over WebRTC, a URL/QR analytics platform with FastAPI + Postgres + Redis, a roofing business site with an embedded ElevenLabs voice booking agent, and learning resources on Docusaurus/GitHub Pages. Owned auth, rate limiting, cost protection, and deploys across Vercel, Render, Neon, Upstash, and Netlify.

Next.jsFastAPIOpenAI RealtimeWebRTCPostgreSQLRedisDocker

B.S. Computer Science

Dec 2022

University of Maryland, College Park

Coursework in Advanced Data Structures, AI, Data Science, Computer Architecture, OOP, Functional Programming, Web Development with JavaScript, and Computer Systems.

Data StructuresAIData ScienceWeb Dev

05. contact

Get In Touch

Whether you have a project in mind, a job opportunity, or just want to chat — my inbox is always open.

Location

Lynnwood, WA · Remote-friendly

Find me on