Key Strategies for Building a System: A Tag Platform
Finding the highest-ROI MVP inside a stalled plan, then layering tags by recall, filtering and display through metadata
I build products that help people find what matters: first search, then AI agents, now investing.
Product manager at ByteDance Fintech, in wealth management. Before that, the agent platform at Baidu and search at Meituan. Six cities, three countries, one football.
Explore my footprints ↓ · See what I built → · Let’s talk ✉
We’re living the future so the present is our past.
Work: AI Agents & MCP Search & Recommendation 0→1 Growth Fintech Data Platforms A/B Testing & Causal Inference
Life: Football captain Outdoors ENFP Six cities, three countries
Updated September 2026.
Search → AI agents → investing. The common thread: helping people find what matters.
Launched a securities & fund investing platform 0→1 on a referral model, and built out the core journey: discovery, watchlist, user reach and an AI investing assistant.
Turned first- and third-party capabilities into standardized MCP servers and Skills for agents, including the Baidu Search Skill.
On top of a complex search engine architecture, our product team made sure users get the best possible results. I bundled the tag services, filter services and tag recall links into one tag data infrastructure, and led an open platform that lets every business team use that data without waiting on engineers.
Investors are still caught out when listed companies release their results. We trained models on third-party datasets to make daily predictions, so investors get a more accurate, real-time sense of how companies are performing.
Coupons are a go-to lever around Black Friday and Cyber Monday, but do they really drive sales? I built templates and automation so business teams could measure coupon impact themselves, and an A/B test analysis tool that earned an eBay Spot Award.
More on what I built at each stop → · Working on something similar? Let’s talk →
2011–2014 The High School Attached to Shandong University, Jinan.
2014–2017 Shandong Experimental High School, Jinan.
2017–2021 Shanghai University of Finance and Economics. Bachelor’s in business analytics, with a second degree in computer science.
2019 Exchange at IE University in Madrid, drawn by its QS top 10 ranking and Spanish football!
2021–2023 MSc Analytics and Management at London Business School (ranked 5th in Business Analytics), with an exchange at HKUST in Hong Kong.
I grew up in the Jinan Iron & Steel community, which my grandparents helped build in 1958. It had its own hospitals, schools, a stadium, and lantern shows that filled the main road. The mills closed in 2017, the same year I left for Shanghai, hoping to help make the sky bluer and the water cleaner through better decisions. Read the short story →
Jinan → Shanghai → Madrid → London → Hong Kong → Beijing. Tap any city on the map above for the story behind it.
Football since 2011. In Shanghai our university team won a quarter-final place on penalties, and dominated a game we still lost. In London I learned that “man on” is not someone called Manon (football English, explained). Off the pitch: all kinds of outdoor sports.
Reach out if you’re building with AI agents, working on search or fintech products, want to swap notes on product careers, or just need one more player for a football game.
Or pick whatever is easiest for you:
| Email Me | GitHub | ||
| Book a coffee chat | |||
| Résumé (PDF) | |||
Finding the highest-ROI MVP inside a stalled plan, then layering tags by recall, filtering and display through metadata
Tableau Visualization
Tableau Visualization
It’s like how to ‘group by’ some categories and run many regressions.
How to load google docs in R
Many kinds of data visualization here.
AM12 revision notes
What’s the difference between these two similar functions?
Let hugo bind your IP address
Met a senior student in MiF graduated from my high school in Jinan, China!!