Results That Speak for Themselves

Major European B2B electronics distributor with legacy search infrastructure and no ML platform.

Legacy search delivering poor product discovery results. No ML infrastructure for experimentation or production deployment.

Built enterprise AI platform on GCP with full MLOps. Deployed AI-powered search, personalization, and recommendation systems at scale.

European retail and e-commerce companies allocating significant multi-channel marketing budgets.

Marketing budget systematically misallocated due to last-click attribution — rewarding the last touchpoint rather than the channels that actually drive purchase intent.

Implemented ProDataAI's conditional probability-based attribution model using all journeys — both converted and non-converted — with multivariate time series for budget forecasting across SEO, SEA, Display, and TV.

Analytics consultancy client with static pricing strategy across a large product catalog.

Static pricing leaving significant revenue on the table — no visibility into how demand responds to price changes across segments and categories.

Built price elasticity model analyzing demand sensitivity across product categories, customer segments, and competitive context. Integrated directly into pricing workflows.

German law enforcement agency requiring faster, more reliable analysis of visual evidence in criminal investigations.

Manual, time-intensive image analysis creating bottlenecks in investigations. Human review of large evidence volumes was neither scalable nor consistent.

Designed and developed "SherlockCV" — an AI-powered computer vision system for automated evidence image analysis, built with the reliability and audit-trail requirements of law enforcement.

Major global technology company — EMEA AI Cloud consulting practice in early stages of development.

Nascent AI consulting practice required rapid scale across Europe — building team, methodology, client base, and delivery capability simultaneously.

Built and led the European AI consulting team from the ground up. Established delivery frameworks, hired and mentored senior consultants, and landed key enterprise accounts across EMEA.

Large enterprise transformation initiative requiring broad AI capability uplift across business units.

Low AI adoption across business units despite significant technology investment. Employees lacked the confidence and skills to integrate AI tools into daily work.

Designed and delivered a structured AI training program with hands-on workshops, role-specific learning paths, and an AI Ambassador certification track to create internal change agents.

Major European B2B e-commerce operator with growing fraud exposure and manual review bottlenecks scaling with transaction volume.

No automated fraud detection layer — manual review teams unable to keep pace with volume. Fraud losses accumulating with no systematic way to prioritize high-risk transactions.

Built production fraud detection AI with real-time transaction scoring, automated review routing, and continuous model retraining pipeline — deployed across 12 European markets.

Major European insurance group with manual claims assessment processes and high liability miscalculation rates across a large book of business.

Claims reserves systematically miscalculated due to manual assessment — creating significant financial exposure. No predictive layer to identify high-risk claims early.

Developed ML-based claims liability optimization model integrating claims history, risk signals, and external data to predict reserve requirements and route high-risk cases for early intervention.

Major European fashion retailer (€500M revenue) offering customer credit — with static credit limit rules leaving significant revenue unrealised.

Blanket credit policies unable to distinguish between risk profiles — over-restricting creditworthy customers while under-flagging high-risk ones. Revenue and risk both mismanaged.

Built ML credit scoring model combining transaction history, behavioural signals, and repayment patterns to dynamically assign credit limits at the individual customer level.