A Forward Deployed Engineer, built for North.

Because FDE is the work I love most, stripped of the agency overhead. As CIO my role drifted into running a client-services business; the part that lights me up is being in the room with a customer, turning an ambiguous problem into a shipped agent. Cohere lets me do exactly that with frontier models behind me and North as the platform — and the enterprise, regulated focus is where my decade of scar tissue is most leveraged. I'd rather help define that frontier from the inside than keep grinding it from the outside.

I won't pretend Python is my deepest production language — it's where I do AI and data work, not my primary service language. But the bar the JD actually sets is clean, testable, observable, scalable code, and that's exactly how I ship in TypeScript and Go: typed contracts, CI gates, Jest/Cypress/Playwright suites, structured logging. Those habits are language-agnostic. The agentic ecosystem is Python-first, so the ramp is fast — I'd be building real North agent features in Python from week one and carrying that discipline across. I'd rather be the engineer who's honest about the gap and closes it visibly than the one who oversells it.

The same way I treat compliance: designed in and measurable from day one, not bolted on. I'd start by pinning down success criteria with the customer, then build golden eval sets and captured traces so every change is graded against accuracy, safety, and latency rather than vibes. Those scorecards go into CI so regressions are caught automatically, and I'd separate offline evals from production telemetry so we see real-world drift. I've run rigorous automated test suites and compliance audits for years; formalising that into agent evals is the natural extension, and it's where I'd invest early.

I've shipped into exactly those constraints — HIPAA EHR pipelines at EKAGRA, GDPR + VAT commerce at Jamahook, federated OIDC and instance-level RBAC at Anygum — so I architect for isolation, least-privilege access, and auditability before the agent ever touches sensitive data. For agents specifically that means scoped tool access, human-in-the-loop on high-stakes actions, full traceability of every step, and evals for safety as a release gate. Enterprise adoption rewards predictability, not the flashiest demo, which is the whole reason a secure, sovereign platform like North is the right place to do this.

Because this is a deployment role, not a research role — the job is taking the lab's models into the messy reality of enterprise, and that's what I've done for ten years. I've already put Claude into a client's production stack and shipped agentic products in public. The team building North doesn't need another model trainer; it needs someone with deployment callouses who can earn an enterprise's trust, scope a use case end-to-end, and ship it reliably. That's the exact shape of my last decade.

Yes — I'm intentional about basing this chapter in London or Europe, and being on-site with customers is the part of the job I want, not a cost I tolerate. I've delivered for London, Zurich, Oslo and Sandnes clients for years and am comfortable in the room with both engineers and executives. I relocate as soon as the visa processes (UK Skilled Worker / EU Blue Card) and commit to local hours and customer travel from day one.