Data scientist.
Builder of things
that work.
A decade turning messy problems into working systems — across experimentation, growth, and AI products. I write about what I learn along the way.
Writing
All posts →Things I've shipped
All projects →About
I’m Kevin Pan — a data scientist turned builder.
I build AI-powered products, from LLM applications to data tools that help people get answers faster. Currently I’m a Senior Data Scientist on WhatsApp’s Growth team at Meta, and I build with Large Language Models on the side, including askdata.app for exploring data through natural language.
Background
My career path has been: data science → product → building. Along the way:
- Growth Data Science at WhatsApp / Meta — Driving growth and integrity strategy, experimentation, and AI data tools for WhatsApp’s Growth team
- AI Product Development at Strivr — Primary AI owner; built a context-aware LLM system for content generation and co-designed an LLM evaluation rubric
- LLM Applications — Building custom content generation and data exploration tools with large language models
- Product Data Science at Lyft — Led experimentation and data modeling on the Integrity and Delivery teams
- Data Science Consulting — Helped organizations like BUILD define their North Star metrics and connect data to impact
- Open Source — Contributed to Delta Analytics’ ML for Good curriculum and built R packages for data cleaning
Consulting
I help organizations build data and AI products — from defining what to measure to shipping working software.
- LLM Application Development — Building AI-powered tools on top of your organization’s data
- Data Strategy — Identifying the right metrics and connecting data to business outcomes
- Prototyping & MVPs — Going from idea to working product quickly
Let’s talk
If you’re building new products — especially at the intersection of data and AI — I’d love to chat. Reach out at hello@kpan.dev.
Work
Helped organizations define North Star metrics, connect data to impact, and contributed to open source ML curriculum.
Led experimentation and data modeling on the Integrity and Delivery teams, focusing on user outcomes and understanding the why behind what we build.
Primary AI owner; built a context-aware LLM system for content generation, co-designed a multi-dimensional LLM evaluation rubric, and owned product development end to end.
Drive growth and integrity strategy, experimentation, and AI data tools for WhatsApp's Growth team, and designed the user identity model now canonical across Growth and Integrity.
Building LLM-powered products on the side, including askdata.app for exploring data through natural language.