I build the machinery that modelling work runs on.
I'm an engineer who does data science. Eight years of shipping software, most of it now pointed at the parts that decide whether a model keeps working: pipelines, orchestration, experiment tracking, evaluation, CI. I work independently, mostly with Havas Media.
Most models work once, in a notebook. My work starts there and ends with something that runs on a schedule, is evaluated the same way every time, and can be rebuilt by whoever inherits it. Above that sits the layer I spend most of my time on: the templates, libraries and infrastructure that turn the next client's build into a configuration job. The problems I look for are the ones where modelling and engineering can't be pulled apart: lifetime value, incrementality, attribution.
Client work isn't mine to publish, so what's here is the thinking behind it. The slip-box is most of that: short notes on statistics, modelling and experiment design, one claim each, each linked to the claims it rests on.
Current client work
- 2026.09 Job orchestration platform rebuilt as a React SPA The internal tool the team uses to define, schedule and run pipeline jobs. The front end was rebuilt in seven phases: a contract snapshot and session hardening first, then the app shell, tasks, logs, jobs, schedules and admin. The new SPA ran alongside the old one until the last phase retired it. The same release closed a data-access gap, so job and instance records are now filtered by ownership.
- 2026.09 A template-and-Terraform framework for client onboarding Every new client meant hand-building the same seven pipeline stages. Each stage became a template repository, and the infrastructure around it moved into Terraform: foundation, transformation, model application, user unification, orchestration and compute quotas. That included the per-client deployment definitions, which had been made by hand and had drifted apart.
Current personal work
- 2024 - 2026 zeszyt-toolset What does it take to turn a paper scoreboard into a prediction system? A notebook of FIFA results kept by hand since 2020, rebuilt as a system. It reads match stats off screenshots, rates players on a custom ELO that accounts for the club they were dealt, draws fixtures balanced on skill and club strength, runs seasons and cups, and predicts scorelines two different ways.