You'd be building AI tools for the people who run Pfizer. The Office of the Chief Executive Officer (OCEO) sets the company's strategy, looks after its communications and global health investments, and runs the CEO's own office. So the people using what you build are the CEO's team and the most senior leaders in the company, and they need it to be right as well as quick.
Most of the job is getting early ideas into production. Some turn up as an idea and nothing more, and you design and build from scratch. Others arrive as a prototype or a pilot that proves the point but isn't secure, reliable or supportable yet, and you can't measure whether it's working. Either way you end up with something colleagues can depend on, built with current AI techniques on Pfizer's enterprise AI platforms.
It's a hands-on Director role in a small, senior team that moves fast and owns its own architecture. You'll work hand in hand with the Director, Product Lead: they decide what's worth building and why, you decide how, and neither job works without the other. You'll also work with two Engineering Associates, and they'll pick up most of their standards from watching how you work. If you want hard technical problems, and you'd like to be part of creating a team culture as well as software, it's a good fit.
Role Responsibilities
Get ideas into production When an idea has earned it, you take it all the way. If nothing's been built yet, you design and build it with the person who came up with it. If there's a prototype or a pilot, you work out what it does well and where it falls down. Then you harden it for real use: security, data handling, testing, monitoring, cost control, and a named owner. If only the original author understands it, it doesn't go live.
Own the architecture You own the technical design of everything the team builds and keep it hanging together, so each new piece adds to a shared set of components instead of sitting on its own. The question you keep asking is whether Pfizer's platform already does this, or whether it's really specific to OCEO.
Write and review code You'll write and review code most days. AI tools do most of the implementation. You keep the architecture and the reviews in your own hands, and knowing where that line sits is a big part of the job.
Refine the agentic software factory The team builds with an agentic software factory. Every piece of work goes through the same steps: specify, plan, build, review, ship. AI agents do most of the building, and engineers check what comes out. You'll keep improving it: tightening the specification and review steps, choosing the tools and agents it runs on, and fixing whatever keeps sending work back for another pass.
Work with the platform teams You're the technical contact for Pfizer's enterprise AI platform teams. You pick up their new capabilities as they land and tell them what OCEO needs. When a platform slips, you change the plan rather than building a shadow version of it.
Get things approved Solutions go through the Architecture Review Board, the Security Review and the AI Review Board, and you help with the privacy and risk assessments. Nothing touches real data or real users until it's approved. Wherever AI output acts in a system or leaves OCEO, a person signs it off. Access stays as narrow as the work allows.
Take over partner work Sometimes external partners or colleagues build something and hand it over. You agree up front what a good handover looks like, check it before you accept it, and fix the gaps. OCEO shouldn't end up supporting something it can't change.
Run it in production Once it's live, you run it: monitoring, evaluation results, cost, incidents, rollback. You should know what each solution costs and what it gives back, and step in early if quality starts to slip or the bill starts to climb.
Build with the people who had the idea Work with the colleagues who bring ideas in, so they know what they're taking on when it goes live. Make it easy for them…