About
I'm a full-stack engineer with 4+ years of production experience, specialising in Go backends, React/Next.js frontends, and DevOps pipelines — and over the past year I've gone deep on AI application development, shipping full-stack AI products end to end.
Recently I've been designing and building a series of AI-assisted products — an AI journal, a book cataloguing app, a community resource explorer, and a relationship manager — each involving real architecture decisions: API design, data modelling, and agentic tool-use pipelines, not just fast prototyping. View my projects on GitHub →
I take ownership of what I build end to end, and I'm a lifelong curious learner — I like sharing what I learn in public, on X and Medium.
Outside of work I love travelling, reading, hiking, and handcrafts like knitting and sewing.
I'm currently looking for a new role where I can bring both sides together — solid full-stack fundamentals and hands-on AI/agentic engineering — on a team building real products.
Experience
Professional Experience
Opusv Technology
- Designed a dual-layer Go backend (external + internal, JWT auth, CORS isolation) for a self-service client portal serving 40+ enterprise clients, replacing manual SAP data exports.
- Built a DNS management system replacing a legacy shared-credential workflow across 100+ client domains — cut support tickets 85% and reached 99.9% update reliability.
- Delivered a RADIUS/VPN session monitoring dashboard for 500+ users with Keycloak OIDC auth; automated CI/CD cut deployment time from 2 hours to 15 minutes.
- Go
- React
- Next.js
- TypeScript
- AWS
- Docker
- OAuth2/OIDC
- Keycloak
Education
Deakin University
Master of Business Analytics (Information Systems)
Volunteer & Community
RMIT STEM Mentoring Program
Industry Mentor
- Providing one-on-one career guidance to a Cloud Computing student over four structured sessions, including a workplace visit.
WaterAid Australia
Database Volunteer
- Supported the fundraising team with data hygiene projects in the Raiser's Edge database.
Skills
Frontend
- React
- Next.js
- TypeScript
- TailwindCSS
- Styled-Components
Backend
- Golang
- REST API
- Node.js
- ExpressJS
- C#/.NET
AI & APIs
- Claude
- OpenAI
- LangGraph
- MCP
- RAG (pgvector/LanceDB)
- Inngest
- HuggingFace Transformers
- Python
- FastAPI
- Notion API
DevOps & Cloud
- Docker
- Nginx
- AWS
- Azure (AD/OIDC, MSAL)
- CI/CD
- Vercel
Datasource
- MySQL
- SQL Server
- Redis
Security & Auth
- OAuth2/OIDC
- CSRF
- JWT
- Keycloak
Projects
A set of AI-powered tools I've designed and shipped over the past year, each solving a real problem in my own life — from journaling to book cataloguing to local community discovery.
AU Council Explorer
Discover community resources across Victorian councils — all on an interactive map.
- Childcare, libraries, hospitals, playgrounds, and local events in one map view
- Natural-language search powered by Claude, backed by real-time web search
- Data pulled from a Postgres/Prisma backend across 79+ local councils
- Next.js
- Anthropic
- Mapbox
- Prisma
- Vercel
hand-log
Turn messy daily notes into structured, meaningful reviews — automatically.
- Voice or text entries transcribed and structured by AI in seconds
- Auto-generated weekly and monthly reviews saved straight to Notion
- AI chat window surfaces patterns across your own past entries
- Next.js
- Anthropic
- Notion
- TypeScript
- Vercel
lovely-shelf
Photograph a book cover — AI catalogues it to Notion in seconds.
- Claude Vision recognises title, author, and genre from a single photo
- Stats dashboard: genre breakdown, reading heatmap, author countries
- Compresses a 5-minute manual cataloguing workflow down to 5 seconds
- Next.js
- Anthropic
- LangGraph
- MCP
- Notion
- Tailwind
- Vercel
Linknest
A bookmark manager that actually finds what you saved — semantic search, not just search-by-title.
- RAG search pipeline (pgvector + OpenAI embeddings) with a cost-aware router that sends simple tasks to cheaper models
- Diagnosed and fixed a silent classification bug in production, restoring accuracy to 99% of bookmarks
- Benchmarked a self-hosted fine-tuned classifier against Claude — 37x faster per call, with a documented accuracy tradeoff
- Next.js
- Anthropic
- OpenAI
- RAG
- pgvector
- Inngest
Hackathons
I actively seek out time-pressured, build-in-public settings — hackathons are where I test how fast I can go from idea to working demo. This is an ongoing part of how I keep learning and staying connected to what's happening in the industry.
Kick WinWin — a gesture-controlled audience interaction platform supporting the full streamer journey (before, during, and after a live stream): AI-generated pre-show topics, real-time hand-gesture interaction (point to subscribe, heart shape to send love, fist to pause) and mini-games during the stream, and a post-stream recap. Built under a 12-hour time constraint with a team of 4.
- Collaboration under pressure: Worked closely with the team under real 12-hour time pressure, navigating scope cuts and technical roadblocks together to ship a working demo.
- Product mindset: Prioritised the features that would actually land in a live demo (gesture interaction, mini-games) over ones that just sounded impressive.
- Smart use of AI tools: Owned pre-show topic research — used the "Last 30 Days" AI skill to pull trending topics from X, YouTube, and Reddit, applying AI tooling deliberately to solve a real content-prep problem rather than as a gimmick.
- Audience interaction design: Built the interactive game segment — a collaborative mini-game for the audience to play together during the stream.
- Presentation & networking: Helped structure and time the team's presentation to judges and industry mentors, and used the Q&A as a chance to network with mentors afterwards.
Blog
I write about AI, building personal tools, and things I find interesting.
I Built an AI Journal That Thinks With Me
I kept writing fragmented notes that never turned into anything useful. So I built HandLog — speak or type freely, and AI structures your day into mood scores, key people, and reflections, then saves everything to Notion automatically.
- AI
- AI-Assisted Dev
- Next.js
- Notion
I Built a Full-Stack AI App in 7 Days as a Beginner
Lovely Shelf started as a weekend project — photograph a book cover and have it instantly catalogued in Notion via Claude's vision API. What began as a personal tool for my overflowing library turned into a full product with dashboards, AI chat, and a quotes collection.
- AI
- AI-Assisted Dev
- Next.js
- Notion
A Practical Guide to MCP (Model Context Protocol)
How it works, why it matters, and how to use it with Claude Code, Notion, Copilot, Cursor & more.
- AI
- MCP
- Claude
8 AI Tools That Are Revolutionizing Creative Content
From text-to-video to real-time drawing and AI-generated podcasts — tools changing the way we create.
- AI
- Tools
Face to Sticker Generator using Notion, Make, and Replicate APIs
From Notion to AI-generated face stickers — fully automated, refreshingly simple.
- Notion
- Replicate
- Automation
Every Developer Should Know Git
Whether you're a senior developer or just starting out — mastering Git commands is crucial.
- Git
- DevOps
Redis in System Design
How to use Redis effectively with Laravel — with CLI, RedisInsight, and best practices.
- Redis
- Laravel
- Backend
Wonderful VSCode Extensions
Incredibly useful VSCode extensions that have significantly improved my workflow.
- VSCode
- Tools
Community
Coffee chats → workshops → volunteer → speaker 🎤 — my AI community journey over the past year, from attending to building in public alongside other engineers to speaking on stage.
- Delivered talk "Turn Messy Days into Meaningful Memories" — how non-technical people can use AI to build real tools that solve their own pain points.
- Used AI Journal as a case study: API integration, product thinking, and what it means to learn by building.
- Strengthened my confidence in public speaking, knowledge sharing, and helping others grow through AI.
Deep dive into Claude Code, Claude Design, and Workflows — everyone brought a laptop and executed their own AI projects, sharing discoveries and debugging together in real time.
Workshop focused on Skills, Hooks, Plugins, and Codex — three hours of hands-on discussion and execution, everyone building in parallel and sharing what worked.
Hands-on workshop — everyone brought a laptop and built for three hours. Topics: MCP, RAG, and Agents. Shared experiments, compared approaches, and debugged side by side.
A relaxed, open discussion — reflecting on how we each used AI over the past year, what surprised us, what didn't work, and what we're planning to build or explore next.
Supported senior speakers who shared real-world AI workflows in full-stack engineering and everyday life. First time seeing how a community event runs behind the scenes.
Deeper discussions on integrating AI into everyday workflows and lowering the barrier for non-technical users.
Continued community conversations around AI tools, side hustle workflows, and comparing models on cost and capability.
First Coffee Chat. Shared experiences using AI in daily life — workflows, side hustles, pitfalls, and which models give the best value.