A high-profile central Government AI unit is looking for senior Forward Deployed Engineers to join a small, fast-moving team shipping real AI products into frontline public services. You will own the full AI and software stack end to end, working directly with operational users to deliver high-impact tools that make a measurable difference, with security and production reliability built in from the start rather than added on afterwards.
This is a unit that sees itself as a disruptor within government. It moves quickly, works in the open with the people who use what it builds, and is looking for engineers with the same mindset.
What you'll do
- Personally own every part of the stack on your projects: frontend, backend, database, AI, observability and monitoring, and containerisation, not a slice of it
- Build and ship AI-native products into production, including LLM and agentic workflows, speech-to-text, translation, document intelligence and data pipelines
- Work directly with frontline operational users in courts, prisons or probation to discover problems, test assumptions and improve what you build from real feedback
- Move between several concurrent projects and problems, keeping delivery quality, security and judgement consistent throughout
- Take products from early prototype through to production deployment and adoption at scale
What you'll need to show (this is where most applicants fall down)
- Genuine, personal, hands-on ownership of the whole stack on real projects. Directing a team or overseeing frontend work someone else built does not meet the bar here
- Real production experience on both sides: a React or Next.js frontend and a Python backend, built by you. Products with only one basic interface are not enough. You should be able to talk through something you have built with real interface complexity, for example separate views for different user types or several distinct workflows in the same product
- Backend architecture depth, ideally with FastAPI, and the ability to walk through how you have implemented authentication end to end, including OAuth and Microsoft Entra ID (Azure AD): app registration, token issuance and validation, and how that protects your API routes
- An applied, practical understanding of how modern AI systems work and fail: context engineering, retrieval-augmented generation, tool-using agents, and the fact that models behave probabilistically rather than deterministically
- Judgement about security and production reliability as part of how you build, not a separate step, and a track record of designing safeguards such as prompt-injection defences and least-privilege access
- Comfort working in ambiguity, moving quickly, and being straightforward about what you built yourself versus what a team around you did
Also useful
- The Next.js, FastAPI and Postgres stack, or something close to it
- Experience delivering technology in a regulated, safety-critical or otherwise high-consequence environment
- Making pragmatic build-versus-buy decisions and integrating third-party models, platforms or data services
- An entrepreneurial or startup background that shows you can deliver with limited structure around you
- Supporting or mentoring other engineers and raising engineering standards across a team
Rate and logistics
Day rate c.£700, inside IR35. You will need to hold, or be eligible for, UK government security clearance (BPSS as a minimum; SC is desirable and will be prioritised).
#J-18808-Ljbffr
