Linux-first Data Engineer focused on distributed data systems, automation, and infrastructure-minded software.
I build practical systems around PySpark, Kafka, Airflow, Docker, and Python: pipelines that can be observed, replayed, debugged, and operated from a terminal without ceremony. My work leans toward production data platforms, developer tooling, TUI interfaces, and clean operational habits.
data pipelines / linux systems / terminal tooling / distributed workflows
| Project | Focus | Engineering angle |
|---|---|---|
| pyspark-medallion-platform | PySpark, lakehouse architecture, medallion layers | Batch data platform with bronze/silver/gold modeling, validation, and reproducible local execution. |
| airflow-dbt-warehouse | Airflow, dbt, warehouse orchestration | Scheduled ELT workflows with explicit dependencies, transformation boundaries, and maintainable DAG design. |
| streaming-kafka-data-quality | Kafka, streaming, data quality | Event-driven validation pipeline for catching malformed records before they poison downstream systems. |
| tui-news-reader | Python, TUI, terminal UX | Keyboard-driven terminal application built for fast reading, filtering, and low-friction daily use. |
| document-quality-pipeline | Automation, parsing, quality checks | Document ingestion and validation workflow focused on repeatability, traceability, and clean outputs. |
| .dotfiles | Linux, shell, editor workflow | Personal operating environment for terminal-first development, automation, and fast context switching. |
Languages
Python, SQL, Bash
Data Engineering
PySpark, Apache Spark, Kafka, Airflow, dbt, ETL/ELT, data quality, medallion architecture
Infrastructure
Linux, Docker, Git, GitHub Actions, local-first development environments, reproducible workflows
Linux Tooling
tmux, Neovim, zsh, shell scripting, dotfiles, CLI automation, TUI applications
Observability
Structured logs, pipeline checks, failure visibility, operational metrics, debuggable data flows
- Building data platforms that are easy to run locally and realistic enough to resemble production.
- Deepening distributed systems fundamentals around streaming, partitioning, retries, and backpressure.
- Improving pipeline observability with better validation, logging, and failure surfaces.
- Sharpening terminal-native workflows with tmux, Neovim, zsh, and small automation tools.
- Turning personal infrastructure into reusable open-source patterns.
My preferred interface is a shell session with sharp tools and low friction.
editor neovim
shell zsh
session tmux
workflow git + make + docker + scripts
style keyboard-first, observable, reproducible
I care about fast feedback loops: commands that explain themselves, logs that point to causes, scripts that can be rerun safely, and local environments that make debugging boring in the best way.
Good systems are not just systems that work once. They are systems that can be understood later.
I optimize for:
- clear data contracts and explicit pipeline boundaries
- scalable designs that respect failure modes
- observability before guesswork
- automation for repeatable operations
- maintainable code over clever code
- small tools that remove daily friction
I use GitHub as a public engineering notebook: experiments, infrastructure patterns, data systems, terminal tools, and the working surface of my Linux environment.
The thread across the repositories is simple: build systems that can be run, inspected, and improved by another engineer.