The full track record.

Independent since April 2026 — building and leading data and AI products for clients, from R&D to shipped, and helping teams do the same.

Currently embedded with the founder of EnergyLabs, taking their energy-optimization engine from R&D toward production: a deterministic, explainable battery-sizing engine built on linear programming (Python, PyPSA, HiGHS). It's exactly the applied-optimization, zero-to-shipped work I like best.

Founded the Intelligence & Automation team from scratch and grew it into a self-directed unit of full-stack data scientists operating at the intersection of product research and production engineering.

The team owned its problem space end-to-end — from identifying user pain points and prototyping solutions to shipping production features — delivering:

• SodaGPT (LLM-powered data quality assistant)
• Time-series anomaly detection
• Automated root-cause analysis
• Semantic data type detection
• Data quality check suggestions

These products became Soda's strongest lever for converting open-source users to paid customers. When it came to integration, the team embedded directly with cloud engineering squads in a tiger-team model — no handoffs, no specs over the wall.

Joined as a Data Scientist in January 2019 and was promoted to Senior Data Scientist in October 2019. Stepped into a young data science function and shaped it into a product-facing operation — leading squads of 2–4 data scientists across multiple product lines simultaneously.

Built and shipped data products with direct user impact:

• Search personalization and content recommendation (Flights & Hotels) — cut content discovery time from 30+ seconds to under 10
• NLP-based receipt itemization for the Liquid expense product
• Revenue and yield optimization (Hotels)
• Geospatial clustering for internal tooling

Worked directly with PMs and C-level to spot product opportunities, scope them, and turn them into production-grade systems — no waiting for a roadmap to tell us what to build.

Operated as a one-person data science generalist across every revenue-critical part of the business — attribution, forecasting, portfolio strategy, and recommendations.

Highlight projects:

• Built a marketing attribution model (Markov chain-based) that directly added +€5M/year in net profitability by reallocating spend to the touchpoints that actually converted
• Created a 12-month demand forecasting system that shifted the sales team from reactive sourcing to proactive planning
• Led a portfolio analysis with a team of analysts — cut the catalogue by 40% while keeping net revenue flat
• Solved the cold-start problem for 40% of the product catalogue by replacing interaction-dependent recommendations with deep neural network representations built from item content and imagery

Led a research project on how the brain anticipates and predicts upcoming words during language comprehension, which led to a published research paper, several conference appearances, and an undergraduate student's thesis on the topic. During the project I also gave a TEDx talk on the work.

Investigated how our brains recognise objects in the real world, and how human language drastically changes that experience — using EEG and brain stimulation. The contributions I made were featured in an influential paper in the Journal of Neuroscience that disrupted long-unquestioned hypotheses and led to 5+ follow-up articles and replication studies by at least two labs across the globe.

Carried out research on a project studying the processing of emotion in language, and supported a lab of 10+ researchers with data analysis and experimental setup.

Responsibilities:

• Behavioural and neurophysiological research design
• Lab management and support
• Data analysis (linear modelling, machine learning, Bayesian statistics, signal processing)
• Training and supervising students
• Scientific writing

In my PhD I set out to revisit and test a hypothesis that had long been discredited due to bad experimental practices. Using EEG, I showed that language influences everyday performance on basic tasks such as object categorisation and visual recognition. I also collaborated with teams of researchers on side-projects looking into language processing and emotional processing.