Nan Fung Group

Senior Associate, Data Science

Drive firm-wide AI implementation, working closely with the Group CEO on AI for the Biotech VC and family-office arms — from autonomous learning loops and an AI investment-agents team to a screening platform that cut underwriting from a day to minutes.

What I worked on

• Work closely with the Group CEO on firm-wide AI strategy for the Biotech VC and family-office arms — building autonomous learning loops and an autonomous AI investment-agents team, and driving emerging-technology direction.
• Built a multi-agent platform with a centralised knowledge base over internal “skills” and documents, used across investment, leasing, and operations teams.
• Shipped a full-cycle AI investment-screening platform that cut asset underwriting from a full day to minutes per deal — now used by the deal team, structurally changing how opportunities are sourced, screened, and pursued.
• Built a reusable agent framework with a pluggable memory system powering all production AI agents, lifting agentic task success rates by 20% and cutting end-to-end build cycles by 70% (weeks → days).
• Owned full-cycle delivery: React / TypeScript front-ends over Python (FastAPI) backends with ingestion-to-reporting pipelines, deployed as human-in-the-loop AI agent SaaS on AWS EKS via GitHub Actions CI.
• Deployed NLP, forecasting, and LLM-based synthetic data generation for valuation signals, comparables, and investment-scenario stress testing.
• Applied CART analysis on campaign data to identify high-potential customer segments, driving a 4× conversion lift.