I work at the intersection of business processes, AI-enabled automation, knowledge management and delivery. My focus is not isolated prototypes, but translating modern AI into workflows that teams can actually use — governance-aware and defensible.
- AI transformation & use-case structuring — from business need to prioritized, feasible use cases
- Agentic AI workflows for document analysis, research and operational automation
- RAG / Graph RAG & knowledge management to ground answers and reduce hallucination
- Process automation combining deterministic (n8n) and agentic components
- Requirements engineering & Business-IT translation — bridging stakeholders, product and dev
- Governance, evaluation & human-in-the-loop for critical decision paths
AI & Agentic
Automation & Integration
Knowledge & Graph
Cloud & Infra
Delivery & Certifications
|
Production-grade FastAPI knowledge-graph RAG service: turns multimodal documents (PDFs, Office files, images, audio) into queryable knowledge graphs with dual-level vector + graph retrieval. Multi-workspace isolation, provider-agnostic routing and a REST API built for n8n workflows and AI agents.
|
Contributed a merged fix (PR #3052) to HKUDS/LightRAG: edge properties were silently dropped on upsert with Apache AGE. Replaced the fragile merge-and-set path with an atomic delete-then-create pattern, verified against Apache AGE 1.5.0 / PostgreSQL 16.
|
🛠️ More coming soon — 1–2 further public repositories are in progress, focused on agentic AI patterns and AI-enabled business processes.
AI Transformation · Autonomous AI Agents · Generative AI · RAG / Graph RAG · Business Process Automation · Bridging Business, Stakeholders & Engineering