Work, from problem to result.

Each project is laid out as problem → approach → result, with AI's role called out separately — the ones you can click into and try come first.

Form: Open-source + live website

NoWorries — Open-source AI desktop assistant

A solo-built desktop-grade autonomous Agent: give it a plain-language instruction, and it plans, picks tools, and executes across multiple steps on its own.

Electron · TypeScript · Python · Function Calling

Learn more

Upstream PR: Merged

OpenClaw open-source contribution and memory-system rebuild

A PR merged upstream; a fork that rebuilds two core modules — memory and context management.

TypeScript · Embedding · SQLite · Context engineering

Learn more

Head-to-head: Under a fixed token budget: single Agent scores 25 vs. Sprout's 100

Sprout — a self-growing multi-agent task tree

An open-source framework, designed solo: every Agent decides for itself whether to split, the tree topology has no depth limit, and recursion is the real thing.

Python · asyncio · litellm · Multi-Agent

Learn more

Real benchmark: 4 real contracts (104–348 pages) run end-to-end

ContractLens — AI review for Australian property contracts

Built with a practicing lawyer: turns a 348-page contract into a one-page report — a 10-stage pipeline, 7 AI analysts in parallel, about $1 a contract.

Next.js · FastAPI · LangGraph · Claude tiered calls (Opus / Sonnet / Haiku)

Learn more

From research to live: 1 day (done on 2026-06-11)

This site: a personal website with a built-in AI twin

A website that answers interviewers' questions on my behalf — the website itself is the proof of AI engineering ability.

Next.js 16 · TypeScript · Tailwind CSS v4 · Vercel AI SDK v6

Learn more

Status: Live in production

Chenxi flower wholesale inventory-and-sales system

A full-stack operations system built solo for a flower wholesaler — live in production, with 200K+ RMB in cumulative sales.

Flask · Feishu Open Platform · H5 · Gunicorn

Learn more

Kaggle ranking: Top 10%

Deep learning in practice: ViT image classification and DQN multi-agent

Kaggle competition Top 10% (validation Acc ≈ 98%) + hands-on multi-agent reinforcement learning.

PyTorch · Transformers · ViT/DeiT · DQN

Learn more