Data Engineering · Analytics · Agentic AI
I'm Johnny Li. I build the data platforms and AI agents that turn billions of signals into decisions.
Data Engineer at Meta, working on measurement infrastructure for the monetization ecosystem. Before that: business intelligence in Amazon's finance org, and data analytics program management on Microsoft's Azure Sphere team. These days I think as much about AI agents and evaluation as I do about pipelines — this blog is where that thinking lands.
- Python
- SQL
- Spark
- Airflow
- dbt
- Trino
- Kafka
- Streaming
- ETL / ELT
- Data modeling
- System design
- Data quality
- AWS
- Kubernetes
- Agentic AI
- LLM eval
- Machine learning
- Automation
- Analytics
- Dashboards
- D3 / SVG
- React / TS
- Tableau
- Power BI
What I do
Data Engineering
Dimensional modeling, ETL/ELT architecture, distributed processing, data quality & governance.
AI & Agents
Agentic systems — tool use, task planning, memory — plus LLM evaluation and AI-assisted tooling.
Analytics & BI
Metric standardization, semantic layers, funnel analytics and visualization people actually use.
Experience
Data Engineer
Meta · Signals Growth
Measurement infrastructure for the monetization ecosystem — pipelines, attribution signals, and quality measurement for internal AI agents.
Business Intelligence Engineer
Amazon · Store Finance Science
Financial data models and ETL for procurement and P&L reporting, plus data-quality monitoring for forecasting.
Data Analytics Program Manager
Microsoft Azure Sphere · via Pactera
ETL pipelines and Power BI reporting for product health and customer support insight, and the report frameworks around them.
M.S. Information Systems
University of Washington
Seattle, WA.
Data Scientist Nanodegree
Udacity
The projects from this program are the 2019–2021 posts in the archive.
Now
- Building agentic systems that investigate and resolve pipeline issues on their own, and figuring out how to evaluate whether they actually work.
- Running a home lab as a testbed for AI agents and data tooling outside a corporate environment.
- Writing here about the decisions behind this work — what I chose, what I rejected, and what I'd do differently.