Data platforms that hold up in production

Cut an inherited proof of concept by 95%, three hours to ten minutes, aggregating 13 dimensions over a 150–300 billion row cartesian product on a single 4-core VPS using granularity-based partitioning, zero-copy handling and SIMD. Built medallion ETL pipelines on a node-local lakehouse feeding a data warehouse, set the team's standards for code quality, testing, CI/CD and gitflow, and mentored developers one-on-one.

Built a Python practice from zero to eight engineers, including the engineering standards and the Agile practice around it. Created tech demos as go-to-market entry points into new industries, and led delivery of client web APIs and an AWS MLOps pipeline for computer vision.

Automated the inflow of qualified, self-matched candidates through job-seeker database and ATS integrations, and automated lead generation across 100+ Belgian job marketing sites, delivered straight into the sales pipeline.

Delivered the international roll-out roadmap for WattsOn, an energy contract management platform, and built its ETL and ELT pipelines with domain-driven referential and exposition layers across CSV, JSON, XML and EDIFACT.

Backend and full-stack delivery for Belgian clients, moving from web applications into data engineering.

Black-box root-cause and performance analysis on shipped, closed-source game engines. Credited fixes acknowledged by the studios concerned.

Led the world's largest open-source Pokémon GO project: 2,800 stars, 1,600 forks, 12M monthly visitors. A high-performance GIS platform visualising real-time geospatial data at scale, crowdfunded for three years, with an official Discord partnership and a 35,000-member community.