Open to opportunities

David Draguta

@daviddraguta

Message

I build production forecasting systems and rigorous machine-learning models.

Germany

Message

What I'm looking for

I'm looking for full-time remote work in the EU where I can own rigorous ML and data products end to end, from problem definition and infrastructure through modelling, validation, and production.

At One Auction View, I owned RetailVision’s data-science function end to end and shipped a Databricks forecasting system scoring tens of millions of US vehicle listings daily. I rebuilt time-to-sale estimation as a survival problem, identified observation-window censoring bias, and validated the inferred labels against filed 10-Q used-unit counts.

I work across Python, SQL, Spark, JAX, and Terraform, with particular depth in survival modelling, Bayesian methods, causal reasoning, and experiment design. I’m direct about what data can support: I measured a near-memoryless predictability ceiling across 45M+ exits, rejected unnecessary model complexity, and pushed for first-party behavioural data where supply-side listings could not identify price elasticity.

I also build from primary sources when needed, including an encoder-decoder Transformer implemented from the original paper in Julia, and have delivered computer-vision work for disaster mitigation.

Experience

Work history, roles, and key accomplishments

Data Scientist (Contract)

One Auction View

Jan 2024 - Dec 2026 (2 years 11 months)

Sole data scientist for RetailVision, OAV's used-car analytics product, owning the full pipeline from problem definition to production. Built and shipped a production forecasting system scoring tens of millions of US vehicle listings daily on Databricks, and diagnosed and fixed a systematic bias in the incumbent sales estimator.

Python SQL Spark Databricks Survival Analysis MLE Data Validation Data Strategy