Skip to content
View shuxiachai's full-sized avatar

Highlights

  • Pro

Block or report shuxiachai

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
shuxiachai/README.md

Hi, I'm Jingbo Chai 👋

AI Agent & LLM Application Engineer

Master of Computer Science @ The University of Sydney · 2027 Graduate

I build evidence-grounded Agent systems and production-oriented LLM applications, focusing on workflow orchestration, RAG, evaluation, guardrails, observability, and failure recovery.

Email · Featured Project · Live Demo


Featured Projects

An evidence-constrained 6-agent workflow for turning research topics or papers into citation-backed commercialization reports and TRL/MRL scorecards.

  • Three evidence agents analyze academic, patent, and market evidence in parallel, followed by a sequential Writer → Reviewer → Scorer workflow.
  • Pydantic guardrails enforce structured outputs, citation integrity, immutable sources, and deterministic scoring.
  • Content-addressed checkpoints passed 30/30 hard-kill recovery cases; a 90-unit topology ablation quantified quality, token, and cost trade-offs.

Stack: Python · CrewAI · FastAPI · Pydantic v2 · OpenTelemetry/OpenInference · Docker

Repository · Live Demo


A local-first Australian bushfire preparedness report system combining deterministic analysis, controlled LLM generation, hybrid RAG, and human review.

  • Built a fail-closed hybrid RAG pipeline with Qdrant, EmbeddingGemma, BM25/RRF, jurisdiction filtering, and prompt-injection boundaries.
  • Achieved Top-8 Recall 100%, MRR 92.2%, and Top-1 85.3%; all 8/8 model regression cases passed governed report gates with zero safety violations.

Stack: Python · Streamlit · EmbeddingGemma · Qdrant · BM25/RRF · GitHub Actions

Repository


A local-first task manager for Genshin Impact multi-account players and service studios, delivered as a Windows desktop application and offline PWA.

  • Implemented cross-platform backup migration with ID/UUID remapping and foreign-key reconnection, plus Windows DPAPI credential protection.
  • Delivered 31 GitHub releases across Windows and PWA runtimes and served dozens of users.

Stack: Python · SQLite · JavaScript · pywebview/WebView2 · IndexedDB · PWA · PyInstaller

Repository


Core Stack

Agent & LLM
LangGraph · CrewAI · LangChain · Multi-Agent Workflows · Tool/Function Calling · Structured Output · Guardrails · Human-in-the-loop

RAG & Evaluation
Embeddings · Qdrant · BM25/RRF · Hybrid Retrieval · Context Engineering · Agent/RAG Evaluation

AI Application Engineering
Python · FastAPI · Pydantic v2 · SQL/SQLite · JavaScript · Docker · GitHub Actions · Linux · OpenTelemetry/OpenInference


Open to 2027 new-grad opportunities in AI Agent, LLM Application, Agent Platform, and AI Backend Engineering.

📧 shuxiachai@163.com

Pinned Loading

  1. academic-commercialization-agent academic-commercialization-agent Public

    Turn any research paper into a commercialization report — 6 AI agents, TRL/MRL scoring, patent landscape, market intelligence, verified citations. DeepSeek / OpenAI / Claude.

    Python 581 53