About

The notebook

This blog started from a simple frustration: the technical information that actually matters — the why behind decisions, the errors that never show up in official docs, the patterns you only see after breaking something in production — usually never gets written down anywhere.

Geek Logbook is my attempt to change that, at least for myself.

I started writing to avoid forgetting. Then I discovered that writing to explain forces you to really understand. Now I write because I find it valuable to leave a record of what the craft looks like from the inside: with its shortcuts, its dead ends, and its moments of clarity.

What you’ll find here

Notes from real work. Not polished tutorials — jottings from someone learning while building things.

Most of it is data engineering, and a few areas keep coming back:

• Data pipelines — Glue, Lambda, Data Factory and Airflow in production, mostly after something broke
• Storage — Parquet, Iceberg, Delta, Hive and Trino, and how a data lake is laid out
• Distributed systems — GFS, Dynamo, HDFS internals, CAP, sharding: the design behind the tools
• Data modeling — SQL from the engine’s side, PostgreSQL internals, dimensional modeling
• Healthcare — claims and value-based care data with Tuva and dbt
• API integration — OAuth, tokens, rate limits and retries against third-party APIs
• Analytics engines — ClickHouse, DuckDB and why they are fast

The topics page shows each one with its posts and where they overlap. Python, Linux and cloud tooling show up everywhere, as the tools that get the work done.

The format

Most posts are short and direct. Some are notes from books I’ve read. Others came from investigating a specific bug. A few are longer analyses — when they exist, they’re worth reading to the end.

If something’s wrong or could be improved, the blog’s code is on GitHub. Pull requests are welcome.

Contact

If you want to talk data, infrastructure, or tech in general, find me on the socials below.