Release Notes: Summer
-
v3.0 Current 2026 – now
- Senior AI Product Manager on an AI Innovation team at Microsoft.
- Exploring a simple question: AI already makes individuals 10x faster, so why don't teams get 10x faster too? Finding what blocks human + agent teams, then building tools to fix it.
-
v2.2 2026
- ML Engineer at NextPathway. Built AI tools that move legacy SQL and ETL systems to the cloud.
-
v2.1 2025
- ML Engineer Intern at Roche. Built a multi-agent coding assistant for clinical trial data scientists: 1,300+ queries from 260+ users in its first 3 weeks.
-
v2.0 Major refactor 2024
- Removed: a comfortable big-tech job.
- Added: MScAC (Master of Science in Applied Computing, AI concentration) at the University of Toronto.
- Breaking change: rebuilt from scratch as a researcher and engineer before returning to product.
-
v1.0 Stable 2021 – 2024
- AI Product Manager at Tencent. Owned an enterprise face recognition product end to end, from customer calls to on-premises deployments. 100K+ daily users, 20M CNY in year one.
-
v0.1 Beta 2015 – 2021
- B.Eng. and M.Eng. in Software Engineering at Sun Yat-sen University. Wrote a lot of code, did ML research, and co-authored a paper at AAAI 2022.
Thinking Out Loud
-
Agents Are Powerful, but Are They Really Speeding Up Our Work?
Agents are generating faster and faster, but scattered context and limited human review capacity are becoming the new bottlenecks to real team productivity gains.
-
Building a Startup Inside Big Tech: From Execution-Focused PM to Exploratory PM
What it feels like to do startup-style exploration inside a big company—and to shift from executing against clear goals to finding direction, testing fast, and winning users over.