Projects

Featured Project

pigeonhole

A wall of mail slots for coding agents running in parallel. Each git worktree gets a pigeonhole; agents post what they are working on and leave each other notes, and read the board at session start. No server, no daemon, no MCP — one POSIX shell script and the filesystem.

ShellAI AgentsOpen Source
A 3x3 wall of mail slots, one holding a note, above a terminal showing the pigeonhole session-start message

table-evaluator

Open-source Python library for evaluating how well synthetic datasets match real data. Measures statistical similarity, distributions, and correlations between real and generated tabular datasets.

Python PyPI Open Source

Building the Evaluation Flywheel

Companion repo for the PyData Amsterdam 2026 talk. A data-analysis agent answers 50 business cases, a jury of LLM judges flags where it disagrees, and a human’s answers to those boundary questions become a pass/fail evaluator.

Python LLMs Evaluation

FakeFynder

Website that checks videos for manipulated areas, built as a proof of concept at the Hackathon for Good in The Hague, 2019.

Python Deepfakes Computer Vision

Master Thesis — Synthetic Tabular Data with GANs

Research on improving Generative Adversarial Networks for tabular data generation and evaluation.

Python GANs Research

OpenClaw

Open-source project at orq.ai for AI agent development and orchestration.

Open Source

Context is King

PyData Amsterdam 2025 project exploring long context windows vs RAG for document understanding.

Python LLMs RAG