Principal Generative AI Software Engineer, Citi
Building production AI products in a tightly regulated bank — from retrieval and orchestration to evaluation, security and the last mile of adoption.
Principal GenAI engineer · engineering leader · product builder
I build AI systems that survive real data, real users and real constraints.
I’m Mark Fullbrook, a principal generative-AI engineer and engineering leader. I work across architecture, evaluation, security and adoption, turning complex ideas into products people can trust and use. Today I focus on production generative AI in financial services. Independently, I’m developing Robin, an AI phone agent for small businesses.
The work
From architecture to adoption.
I’ve spent my career designing systems, leading teams and turning difficult technology into working products. Today that means building production generative AI in a regulated environment.
Engineering leader
Led cross-functional engineering teams across London, New York, Tel Aviv and Shanghai. Grew organisations, shipped platforms and stayed close enough to the code to know where the sharp edges were.
Director, family business
Helped run a 65-person company and modernised the systems underneath it. A practical education in operations, people, cash flow and why software only matters when it changes the work.
Independent project
Robin answers the phone when a small business can’t.
Robin is an AI phone agent for owner-operated businesses. It answers calls, understands what the caller needs, captures the right information and makes the next step clear. I’m currently validating it with small-business owners.
Visit RobinHow I think
The hard part starts after the model works.
Make it real
A useful system in production teaches more than a perfect diagram.
Earn trust
AI products need evidence, guardrails and a clear answer to ‘why should I rely on this?’
Stay close
The best technical leaders understand the code, the customer and the constraint.