Antonis Stellas is a freelance data scientist currently working at Nanometrisis, a startup focused on providing software for nanoproduct inspection. His path into data science runs through applied mathematics and physics, a professional doctorate, and hands-on industry consultancy, a combination that lets him move comfortably between rigorous modeling and messy business realities.

Years in startups shaped how Antonis works: cross-functional roles, lean build-measure-learn cycles, and the communication and business know-how needed to ship things that matter. On the engineering side, he works with concrete MLOps tooling and patterns, including MLflow for experiment tracking, Prefect for orchestration, and Grafana for observability, along with model monitoring essentials such as data drift and concept drift, for which he uses and teaches Evidently AI. His project experience ranges from semiconductor prediction to streaming pipelines that push YouTube metrics into BigQuery and Looker, with a strong emphasis on Kafka and Confluent for real-time data.

In addition, Antonis recently joined Upwork, where he offers a range of data solutions to clients. He is candid about what it takes to succeed independently: building a portfolio, iterating on proposals, pricing strategy, onboarding workflows, and invoicing, all while balancing startup commitments. For teams and clients, that translates into a data scientist who understands both the models and the business they serve.