About
I’m Vladimir. Ten years turning mixed-source data into business cases adopted by leadership. IC data-science work at Roxar, Wintershall, and Baker Hughes, then product ownership at Baker Hughes and Wintershall, then two years independent under Rhodaris running JTBD discovery, experimentation frameworks, and LLM+RAG pipelines for client teams.
How I work
Problem framing and hypothesis mapping first. Uncertainty quantified with a stated confidence interval, not hidden inside a point estimate. Business cases sized in the currency the audience uses.
Toolbox
ML & statistics (XGBoost, time-series, uncertainty quantification, causal methods). JTBD, segmentation, TAM/SAM/SOM, experimentation frameworks. LLM+RAG, golden-set evals, guardrails, HITL. Python, SQL, PySpark, FastAPI.
Domains
Energy & subsurface (Roxar, Wintershall, Baker Hughes). Not-for-profit (Giro555 Dutch fundraising alliance, SLK Kosovo reforestation). AI & SaaS (Bevra).
Case studies below. Get in touch if any of them is worth comparing notes on.