Henry Tsai
Legal name: Sheng-Wei Tsai
Full Stack Developer
Professional Summary
Full Stack Developer with a Master of Computer Science from QUT and a strong track record of building real-world applications independently. Proficient in TypeScript, React, Next.js, Python, and PostgreSQL, with hands-on experience integrating AI APIs (OpenAI, Anthropic) into production systems. Passionate about building tools that solve real problems — from AI-powered job search engines to mobile apps. Actively seeking a graduate or junior developer role in Brisbane or remote.
Technical Skills
Projects
Full-stack personal platform combining a technical blog, AI-powered weekly digest, Australian IT job search engine, AI cover letter generator, and job application tracker.
- Built an AI digest pipeline that fetches content from OpenAI, Google AI, DeepMind, HuggingFace, and ArXiv, applies a GPT quality filter, then auto-generates summarised MDX posts with Mermaid diagrams
- Integrated Adzuna API to aggregate thousands of Australian IT jobs with location/type filters, save-to-Supabase, and one-click cover letter generation
- Built a streaming AI cover letter generator (OpenAI GPT-4o-mini) tailored to Australian HR standards, with history saved per user in Supabase
- Implemented GitHub OAuth via Supabase Auth, user dashboard, and application status tracker (Applied → Interview → Offer)
MovieFlix — Mobile Movie Discovery App
GitHub ↗Cross-platform mobile app for discovering and tracking movies, featuring real-time search with debouncing, a custom trending algorithm, and a light/dark theme.
- Integrated TMDB API for live movie data and built a trending algorithm tracking user search patterns in Appwrite
- Implemented real-time search with debouncing to minimise API calls and improve UX
- Delivered consistent UI across iOS and Android using NativeWind (Tailwind CSS for React Native)
AI-Powered Application Tracking System
GitHub ↗Full-stack job application tracker with server-side rendering, AI-assisted features, and containerised deployment targeting cloud platforms.
- Built with React Router v7 using SSR, hot module replacement, and optimised asset bundling
- Containerised with Docker for deployment to AWS ECS, Google Cloud Run, and Azure
- Designed a clean TypeScript-first architecture with 90%+ type coverage
AI Agent in Python
GitHub ↗Autonomous AI agent with function-calling capabilities, able to execute Python scripts, perform file operations, and run a calculator — all orchestrated by an LLM.
- Implemented OpenAI function calling to let the LLM decide which tools to invoke
- Built a modular architecture with separate modules for prompts, function registry, and config
- Designed the agent loop to handle multi-step reasoning and tool chaining
Webbit — Reddit Clone
GitHub ↗A near full-featured Reddit clone with communities, posts (text/image/video/link), comments, upvotes, user messaging, and subscriptions.
- Built core Reddit features from scratch: communities, nested comments, voting, and user subscriptions
- Implemented direct messaging and real-time chat between users
- Designed relational data models for Users, Submissions, Communities, and Comments in Rails
Education
Master of Computer Science
Queensland University of Technology (QUT) · Brisbane, QLD
Relevant coursework: Algorithms & Data Structures, Web Application Development, Machine Learning, Software Architecture, Cloud Computing
Languages
🎯 Match This Job
Paste a job description — AI scores your resume against it and tells you exactly what to improve before applying.