Hi, I'm David!

I build data pipelines and data products, and the applications that put them to use. A background in backend and platform engineering shapes how I build both.

I care about how software behaves in production, not just how it looks in a demo. My background includes platform-facing customer work at Vercel, where I helped customers troubleshoot deployments, routing, DNS, environment configuration, and the kinds of operational issues that appear once real systems are live.

Alongside that experience, I’ve spent years building and self-hosting projects focused on observability, telemetry, infrastructure, and backend architecture. I’m especially interested in the operational layers that keep systems reliable as complexity, traffic, and data scale.

Today my focus is data engineering: understanding messy source systems, modeling them into stable datasets, and building the pipelines and applications that make that data usable. The backend and platform engineering background still shows up in how I work. It's why I care about idempotency, clear contracts between layers, and systems that stay reliable as data and usage grow. Explore my projects, technical writing, or GitHub.

Stack

What I build with

Data pipelines and data models built with Python, SQL, and Databricks, plus the Next.js/TypeScript applications and infrastructure that put that data to use.

Data Engineering

Python
PySpark
SQL
Databricks
Delta Lake

Data Modeling

PostgreSQL
ETL / ELT
Medallion Architecture
Schema Design

Backend & Apps

TypeScript
Next.js
Node.js
Docker

Engineering Principles

What guides my engineering decisions

Understand the source first

Know a source system's grain, keys, and change behavior before transforming anything downstream.

Observability

Data quality checks and pipeline telemetry that surface broken assumptions before they reach a downstream dataset.

Ownership

Own the path from source system to usable dataset: ingestion, modeling, and the contracts downstream work depends on.

Reliable by design

Idempotent, recoverable pipelines. Failures are expected, and recovery shouldn't mean reprocessing by hand under pressure.