Building intelligent systems with a quiet rhythm.
AI engineer in Shenzhen, focused on LLM applications, agent workflows, multimodal generation, and RAG systems. I care about turning prompt experiments into reusable, observable, product-grade systems.
Built a staged LLM pipeline for script parsing, narrative slicing, scene analysis, and shot generation. The system moves storyboard generation from single-prompt output into a reusable workflow with cross-episode character and scene memory.
Reframed the bottleneck from isolated highlight detection to cross-episode identity stability. The rebuilt path supports entity alignment, retrieval, scheduling, and cache reuse for more continuous generated edits.
Designed hierarchical document chunking, multi-stage retrieval, dynamic threshold filtering, and prompt-assisted safety review. Internal tests reached about 90% recall and answer accuracy.
Claude Code plugin with an always-visible virtual pet companion, native statusline reactions, and optional tmux ASCII sidebar.
Development activity recorder for Claude Code: daily notes, architecture docs, changelogs, and weekly outlines.
Sandbox runtime for multi-agent orchestration, focused on structured execution boundaries for cooperating agents.
AI-focused Hacker News digest that fetches and analyzes daily AI and agent-related posts with commentary.
Structured AI-assisted development sprint workflow for planning, execution, and review in agentic coding sessions.
Multi-Agent Contract Protocol for structured file-based coordination between coding agents.
Junlin Zhu, Jiaye Wu, Xudong Luo*, Jie Liu (*Corresponding authors)
Artificial Intelligence and Law (Artif.Intellig.& L), Vol. 32. · Journal Article
Jiaye Wu, Xudong Luo* (*Corresponding authors)
Knowledge Science, Engineering and Management (KSEM), 2021. · Conference paper