Lead Decision & AI Scientist in Brighton

Department: Data and AnalyticsBased in: Brighton HQWorking arrangement: Full time, permanent, hybridAbout MPB.MPB is the world's largest platform for used photo and video kit. We make buying, selling and trading cameras and lenses simple, sustainable and trustworthy, and we're still growing across the UK, US and Europe.Help us build MPB’s AI Automation Hub.At MPB, teams across the business are already using AI to work smarter. Now we’re building a dedicated AI Automation Hub to turn that energy into production-grade, agentic automation that makes a real difference to how we work.The Hub is a new, purpose-built team with its own mandate and backlog. Its job is to identify and deliver high-value internal AI use cases, and to act as a centre of excellence for emerging technology, working closely with colleagues across every function. Where we learn something new, we’ll share it widely.Three roles, one team:The Hub is built around three lead roles that work in lockstep, each part of its own specialist function:Lead AI Product Manager: owns the backlog, prioritisation and lifecycle of Hub work, making sure we focus on the use cases that matter mostLead AI Automation Engineer: designs and ships agentic AI in production, covering LLM orchestration, agent frameworks and robust pipelines, aligned with MPB’s agentic reference architectureLead Decision & AI Scientist: validates opportunities with data, builds and optimises agentic models and data products, and ensures AI capabilities are properly evaluated before they reach productionTogether, you’ll take ideas from first concept to trusted, production-ready tools, with humans kept firmly in the loop.This role.This hub will be our internal centre of excellence for agentic AI, and we need someone to be its data and decision brain.You'll take an exciting idea, such as a new agentic workflow or an internal automation, and turn it into a sharp, time-boxed experiment. Then you'll build it, evaluate it and show what it's worth, with working prototypes rather than slide decks. You'll create the evaluation sets, scorecards and guardrails that make AI systems trustworthy, and build the data foundations and retrieval that underpin them.You'll work as a peer with the rest of the Hub, with no traditional leader and plenty of shared ownership. You'll tell the honest story of every experiment, recommending whether to kill it, double down or hand it off. And you'll coach teams across MPB to innovate with AI responsibly, while keeping a close eye on the frontier.What you'll bring.Substantial hands-on experience in data science, ML or applied AIExperience with agentic workflows, multi-agent systems or production LLM applicationsEvaluation engineering: eval sets, scorecards, regression tests and guardrails for non-deterministic systemsStrong, production-quality Python and the modern AI toolkit (LLM frameworks, vector stores, evaluation tooling)Cloud data experience, including BigQuery and at least one production ML or AI platform (such as Vertex AI)The ability to explain findings clearly to exec-level audiences and to coach non-technical teamsNice to have:Computer vision, multi-modal or retrieval experienceExperience in marketplace, e-commerce or content-rich product environmentsOpen-source contributions or published applied workA Master's or PhD (nice, but we value demonstrated craft over credentials)Why MPB; Why now?Build something new: you'll help define a brand-new function, with real ownership and real impact.Interesting data: kit imagery, content and search give you real problems to solve, in a business that's good for people and the planet.Low-hierarchy, high-trust: the Hub is peer-led and your ideas will be heard.