I’m Lora, an AI Agent Engineer building systems that turn prompts into reliable, observable, and useful outcomes. My work sits at the intersection of multi-agent software engineering, LLM infrastructure, and developer tools.
Build the tools that build the agents — then ship them.
AI Agents · Multi-Agent Systems · LLM Inference · Agent Observability · Developer Experience
I work across a connected set of repositories rather than a single flagship. These are the projects and tracks currently carrying the most momentum.
| Track | Active work |
|---|---|
| Agent engineering | 🔥 ai-engineering-harness · 🔎 Glassbox-Agent-Harness · 🚧 ReposTwin |
| Agent products & knowledge | 🏢 aicompanyos · 🤝 JonLinker · 🧠 second-brain |
| Inference education | 📚 nano-vllm-interactive-guide · nano-sglang-interactive-guide |
| Skills & media tooling | 👁️ free-vision-skill · 📦 packify-skill · hermes-minimax-media + hermes-stepfun-imagegen |
| Web3 & experiments | 🐉 monadmon — on-chain creature raising and PvP, with the economy model in progress. |
| Developer presence | 🌐 loraSys · 🧱 Init-codebase |
Public collaboration is part of the portfolio. Below are repositories I have contributed to through commits or pull requests, with recent merged work called out where it is verified.
| Project | Contribution |
|---|---|
| nishuzumi/moss | Merged: PR #29, an ABI-fetch workflow for explorer-tier tooling; and PR #24, a PancakeSwap V3 single-hop swap adapter. |
| cloudwego/eino-examples | Contributed to framework examples for the Eino ecosystem. |
| datawhalechina/vibe-blog · whale-whisper | Contributions across multi-agent technical-content and collaboration projects. |
| public-apis/public-apis · aymericzip/intlayer | Public ecosystem contributions in API discovery and internationalization tooling. |
| portdeveloper/nad-agent · Ripped-sys/AgentArena-hackthon | Agent and Monad/Web3 collaboration projects. |
Tools I use to design, build, evaluate, and ship agentic systems.
Languages
Agents & LLMs
Product & Infrastructure
A living portfolio across
lora-sysandACAMLab: multi-agent engineering, LLM inference, developer tools, and interactive products.
Agents · Skills · Learning · Products
Six projects that best represent what I’m building now: agentic engineering, observable systems, LLM inference, and reusable developer skills.
| Project | Focus | What it ships |
|---|---|---|
| ai-engineering-harness | Multi-Agent Engineering |
An evidence-driven software-engineering loop with 18 agents and 9 workflows. |
| Glassbox-Agent-Harness | Agent Observability |
Trace actions, inspect context, and evaluate every optimization with evidence. |
| free-vision-skill | Agent Skill |
Compiles images into compact visual-evidence packets for text-first coding agents. |
| nano-vllm-interactive-guide | LLM Inference |
Interactive experiments and Chinese chapters that explain inference-engine fundamentals. |
| nano-sglang-interactive-guide | LLM Inference |
A practical interactive guide to SGLang concepts and serving workflows. |
| packify-skill | Creative Tooling |
Transforms a supplied image into a commercially believable physical collectible concept. |
🔒 ReposTwin (ACAMLab) · a 3D GitHub code galaxy with an AI-agent sidebar · actively in development.
The complete map of self-built and collaborative work, organized by domain. This is intentionally visible: it shows the breadth behind the featured projects.
| Project | Stack | What |
|---|---|---|
| ai-engineering-harness | |
Evidence-driven multi-agent software-engineering harness |
| Glassbox-Agent-Harness | |
Agent observability and evaluation laboratory |
| NanoAgent | |
Customisable agent framework |
| Nanoagent-go | |
NanoAgent ported to Go |
| git-opilot | |
AI CLI for local Git repos · review, refactor, commit |
| Pixel-Heart-OS- | |
Emergent social-universe sim · AI co-editing + Git-style memory (Beads) |
| hermes-stepfun-imagegen | |
Hermes Agent backend · StepFun image generation (step-image-edit-2 / step-2x-large) |
| WishLive |  |
Live-music ecosystem · AI Agent + Web3 collaboration platform |
| Project | Stack | What |
|---|---|---|
| monadmon | |
Monad 上链上养成 + PvP 对战：first living creatures on Monad |
| Newtube-clone | |
YouTube 风视频站 · AI 推荐 |
| Project | Stack | What |
|---|---|---|
| second-brain | |
Personal second-brain dashboard · Obsidian vault sync |
| loraSys | |
Personal site · GitHub Pages (active) |
| SCIDRAFT.AI | |
Science-paper writing platform |
| TrandingOs | |
Personal trading operating system |
| Lora-website | |
Earlier personal website (legacy) |
| Hackthon | |
ETH Beijing hackathon entry |
| demo_monad_hackthon | |
Monad hackathon demo |
| meme-Vibe-Casino | |
Vibe-coding meme casino · hackathon project |
| RecruitDraft | |
LinkedIn job-description generator with star-rule |
| ResumeGenerator |  |
KaTeX / LaTeX resume generator |
| ThreeJs-projects | |
Interactive Sci-Fi control interface |
| lora-bi | |
Personal BI dashboard |
| mianshihou-interview-website |  |
Interview-prep website with backend |
| automatic_compliments | |
自动夸人小助手 · auto compliment generator |
| Project | Stack | What |
|---|---|---|
| redis-go | |
Redis client/server · RESP protocol + data persistence |
| Filehub | |
Local private-cloud file storage |
| node-base | |
AI workflow engine · n8n-style |
| Project | Stack | What |
|---|---|---|
| free-vision-skill |  |
Low-token visual-evidence compiler for coding agents |
| packify-skill | |
Image-to-physical-collectible ideation skill |
| Init-codebase | |
Bootstrap skills for agent coding work |
| ui-design | |
Claude UI/UX design skill |
| web3-frontend-ui-and-ux-reviewer | |
Web3 frontend UI/UX review skill |
| lora-skills | |
Personal Claude skills distillation |
| development-rules |  |
Shared dev rules & workflows |
| Project | Stack | What |
|---|---|---|
| Defect-Supervised | |
Supervised defect detection |
| Defect-Semi-supervised | |
Semi-supervised defect detection |
| DIy-Model | |
DIY-model training playground |
| demo-repository |  |
GitHub showcase template |
| Project | Stack | What |
|---|---|---|
| nano-vllm-interactive-guide ⭐23 | |
13 interactive experiments and Chinese chapters on LLM inference engines |
| nano-sglang-interactive-guide | |
Interactive guide to SGLang inference concepts |
| nanochat-studay ⭐4 | |
Nanochat concept study · 4★ |
| mianshiya-next-lora ⭐3 | |
Interview-prep platform · study project · 3★ |
| neetcode-submissions | |
NeetCode.io problem submissions |
| Snake |  |
Classic snake game |
📦 Earlier experiments · WIP · Private (27 more)
Projects without GitHub description or in early iteration. Listed for completeness; will be either promoted or archived.
AI / Agent helpers
- hackthon-agent — Hackathon-tuned AI agent
- Dream-prop-generator — Dream-prop generator
- skills-manager — Skills dashboard
- aicompanyos — AI company OS concept
Dev tools & apps
- Daily-Rss — Daily RSS reader
- fastgithub — GitHub utilities
- go-project — Go learning project
- better-auth-learn — Better-auth learning
- note-app — Note-taking app
- zchat-demo — Chat UI demo
Skills & automation
- ui-aesthetic-improve — UI polish workflow
- humanize-write — Humanize AI text
Practice & mini-projects
- REACT-PRACEISE — Early React practice
- React-Projects — React practice
- MIscroservice — Microservice practice (typo name)
- demo — Generic demo
- learn-toys — Toys sandbox
- cpp-projects — C++ practice
- cs50 — CS50 coursework
- mode-defect — Defect-detection experiments
lora-sys private: AgentArena · obsidian-notes · agent-projects · nextjs-ai-chatbot · MulitiModal · diy-world-model · CodeTune
ACAMLab private: ReposTwin (the showcase of this profile) · Pixel-Heart-OS- · git-opilot · Defect-Supervised · Defect-Semi-supervised · DIy-Model · demo-repository
Think clearly. Build deliberately. Measure honestly. Ship relentlessly.
I turn ideas into autonomous systems and intelligent workflows — with evidence, not vibes.
Thanks for visiting — the next useful thing is already in the works.