Take client models out of notebooks and into services that stay up and stay accurate.

Zandloper is hired when a proof of concept needs to become something the business can rely on. That is the whole job description, and it is more engineering than research. You will package models, build the serving and monitoring around them, and set up the retraining loop. Python, MLflow, containers, and whichever cloud the client is already committed to. We ask for production experience, not paper experience: at least one model you built that someone is still using. Utrecht, hybrid, and a research day every second Friday that is actually protected.

Machine Learning Engineer