Tomáš Drietomský

We are the first team at Mastercard to build an AI-powered, API-based product on Databricks—GAIME. GAIME is a fuzzy-matching algorithm, which matches transaction/merchant descriptors with Mastercard's real-time merchant database. GAIME consists of entity recognition, candidates retrieval via embeddings + vector search and uses AI-as-a-judge to produce a final match decision. The talk is organized around three foundational pillars of an AI-powered product: governance, responsible AI and quality control. We will discuss how these pillars become accelerators rather than constraints at the regulated financial sector. Finally, we will share the journey of moving GAIME from POC to production and demonstrate what role data science plays in scaling high-impact AI products within a regulated environment.

Our project demonstrates building enterprise AI systems cost-effectively, focusing on matching merchant descriptors to known businesses. Using fine-tuned LLMs and advanced search, we created a solution rivaling alternatives at minimal cost. The system works in three steps: A fine-tuned Llama 3 8B model parses merchant descriptors into standardized components. A hybrid search system uses these components to find candidate matches in our database. A Llama 3 70B model then evaluates top candidates, with an AI judge reviewing results for hallucination. We achieved a 400% latency improvement while maintaining accuracy and keeping costs low and each fine-tuning round cost hundreds of dollars. Through careful optimization and simple architecture for a balance between cost, speed and accuracy, we show that small teams with modest budgets can tackle complex problems effectively using this technology. We share key insights on prompt engineering, fine-tuning and cost and latency management.