About this role
Salary: £42,000 - 82,000 per year
Requirements: 6+ years in software, data, or platform engineering or architecture roles, including meaningful hands-on time on production AI systemsPractical experience choosing between closed frontier models and open-weight/self-hosted models for real workloadsWorking knowledge of AI unit economics, including token pricing, routing, caching, and context engineering, sufficient to model cost credibly at design timeGrounding in our actual stack, including Amazon Web Services AI services and agentic, multi-model patternsComfortable working with regulated, personal-data-heavy employment and compliance data, and treating that as a design constraint from day oneWorking knowledge of information security standards and data protection law as they apply specifically to AI systems, including prompt injection, model provider risk, and data boundaries Responsibilities: Own the design patterns we build our AI systems to, including retrieval and grounding approaches, agent orchestration patterns, tool-calling conventions, prompt construction, guardrail design, and routing logicKeep our AI standards current and document decisions in architecture decision records that Engineering and Product can build againstHold a documented, evidence-based framework for choosing between closed frontier models and open-weight models, and apply it per workloadUnderstand and actively manage the unit economics of every AI solution we recommend, including token cost modelling, model routing, prompt and context engineering, caching, and observability of cost-per-outcomeAssess whether the data behind any AI initiative is structured, accessible, and reliable enough to support the intended behaviour, and flag readiness issues earlyHold the approval path for new AI architectural patterns and material departures from existing standardsAttend initiative design conversations early enough to shape them and sign off on the AI design elements of partner Statements of WorkDefine scope and success criteria for AI spikes and proofs of concept, validate vendor and partner capability claims, and evaluate outputs against real delivery constraintsAct as the internal architectural counterpart to our delivery partners by reviewing proposals and Statements of Work, running design reviews during delivery, and challenging technical decisions where neededSet the standard for how AI systems are monitored and evaluated once live, including offline and online evaluation, drift detection, guardrails, and production incident responseBring Security in as a standing stakeholder on every significant AI initiative and maintain alignment with ISO 27001 and data protection lawDefine and maintain guardrails on what AI systems should and shouldnt decide autonomously, ensure outputs are explainable where they affect clients or employees, assess initiatives for bias risk, and track emerging responsible-AI regulationRun informal sessions or working groups to build AI literacy across Engineering, Product, and the BusinessBe a first point of contact for AI questions and support onboarding of technical staff into our AI standardsMonitor the AI landscape, including the open-weight ecosystem, and evaluate new tools, models, and platforms against our actual needs before they gain internal momentum Technologies: AIArchitectEmbeddedSupportSecurityWebAPIExposedFine-tuningSalesforceSnowflake More:
We are Citation, and this AI Architect role sits within our AI/Data architecture area, embedded in the delivery area it serves while line-managing centrally to our Chief AI Officer. We need a dedicated architectural owner because AI decisions now cut across Product, Engineering, Security, Infrastructure, and the Business, and we want our initiatives to start well and stay on track architecturally. This role formalises work that is already happening informally across our AI delivery, giving it the standing and scope it needs. The role sits alongside our wider architecture function and supports our architecture hub, which owns target-state and standards across the function. We work across a five-layer AI platform architecture and expect this role to help shape standards, governance, and strategic direction across our portfolio.
last updated 37 week of 2026