Things I've shipped

Ships.

Open-source frameworks, production-grade data pipelines, and architectural experiments - each one a problem I wanted to solve.

Framework 01

LogCrest

Zero-dependency async logging for Python with end-to-end flow tracking, a pure-ASGI middleware for FastAPI, and decorator-driven correlation IDs. Production observability without the overhead. v3.0.0 on PyPI.

Python Async ASGI PyPI
Streaming 02

Beyond The Bias

A stream-based news aggregator on Kafka & Flink, scraping RSS feeds and using LLMs for semantic analysis and summarization.

Kafka Flink OpenAI
Ingestion 03

Audible Scraper

Scraping infrastructure on Selenium & BS4 to harvest and clean large-scale datasets for analytical research.

Python Selenium
AI Tooling 04

Plyn

Open-source, local-first AI kanban. Raw ideas come in; your chosen model refines them into briefs with next actions. DuckDB, FastAPI, five LLM providers.

Python DuckDB FastAPI Docker
Developer Tools 05

slot

Open-source CLI that runs AI agents in isolated Docker microVMs - a fresh clone per slot, so agents never touch your real repo. Default mode for speed, safe mode for production.

Python Docker CLI

Smaller experiments

Built for the fun of it.

Want the full impact story?

See how these skills translated into 99.8% reliability and $250K in annual savings.

Back to Overview →