At Netguru, I own the modelling workflow for a retail demand-planning product, turning commercial questions into Python analyses, measurable hypotheses, and decision criteria. I also built a feature pipeline for inconsistent customer and transaction records, raising usable records from 82% to 94% across monthly reporting.
Across Netguru, Netlight, Aiven, Artefact, Hugging Face, and Celonis, I’ve worked on production data systems and machine-learning products, from experimentation frameworks to telemetry pipelines processing roughly 4 million daily events. I use Python, SQL, statistical modelling, and machine learning to test assumptions and make analysis actionable.

