Senior Product Analyst; Artificial Intelligence

Requirements

• You're naturally curious. When you see messy or imperfect data, you want to understand why. You’re also curious about modern AI and LLM tools and have experimented with using building things with them, at work or as personal projects
• Comfortable being given a broad problem and figuring out how to structure it
• You enjoy forming hypotheses, writing code to test them, and interrogating results deeply
• Strong SQL skills and ideally experience using Python for exploratory analysis or prototyping
• You’ve worked in a data-led, fast-paced environment (e.g. fintech, banking, start-up, consulting)
• Ability to communicate clearly and can explain complex findings to both technical and non-technical audiences
• You’re collaborative but independent — you don’t wait to be told the next step

• We’re looking for a Senior Product Analyst to join our AI tribe. This is a builder-style role focused on improving how transaction data is enriched, structured and trusted - and on raising the bar for how we experiment and learn
• This role is about owning ambiguous problems, designing thoughtful experiments, and pushing systems forward through curiosity and disciplined analysis of the outcomes
• Explore the current enrichment pipeline to understand how it works and where it fails
• Form hypotheses about where accuracy is breaking down (e.g. merchant ambiguity, edge cases, new patterns)
• Write code (typically SQL and/or Python) to analyse performance, quantify what's working, and identify systematic gaps
• Build quick prototypes or enrichment experiments to test alternative approaches
• Run structured evaluations, interpret results carefully, and communicate trade-offs clearly
• Work closely with Product and Engineering to embed improvements into long-term system design
• Use AI and LLM tools pragmatically - for exploration, pattern detection, prototyping logic and accelerating experimentation
• Continuously look for deeper patterns and structural improvements rather than settling for surface-level fixes