Lead AI Engineer

Job Title:

Lead AI Engineer Division:
Data, Technology and Innovation Department: AI Product Delivery Salary:
National (Edinburgh and Leeds) ranging from 74,900 to 108,000 and London from 82,300 to 120,000 per annum (salary offered will be based on skills and experience)
This role is graded as:
Technical Specialist – Regulatory Your external recruitment contact is Benjamin via  internal recruitment contact is Lauren via Lauren.

Pyrah.uk Applications must be submitted through our online portal. Applications sent via social media or email will not be accepted.

About the FCA and team

We regulate financial services firms in the UK, to keep financial markets fair, thriving and effective. By joining us, you’ll play a key part in protecting consumers, driving economic growth, and shaping the future of UK finance services. The Data, Technology and Innovation (DTI) division enables the FCA to be a digital-first, data-led smart regulator by delivering a secure, agile, and cost-effective technology and data ecosystem that drives better decisions, transparency, and operational efficiency.

Working alongside the wider AI Programme (which will continue to oversee/coordinate AI activity across the FCA), the department will partner with business leads to shape and deliver work in priority areas — Authorisations, SPC, EMO and Anti‑Money Laundering. This is a hands-on technical lead role building the FCA's AI Platform. You'll design and build, from the ground up, the core capabilities that every AI product at the FCA relies on: retrieval, memory, agent orchestration, model access, evaluation, guardrails and observability.

You'll own the platform codebase and its ongoing development, establishing the technical vision, engineering standards, and quality benchmarks for the team.

Role responsibilities

Lead the AI Platform build and own the platform architecture. Design and run the FCA's AI Platform on AWS as secure, reusable, self-service capabilities with clear APIs, SDKs and documentation. Set the architecture, roadmap and reference architectures, translate constraints across product, data, security, legal and operations into clear technical standards, take designs through enterprise review, and make evidence-based build-vs-buy and model decisions

Deliver a model gateway and access layer. Provide governed access to Amazon Bedrock and other approved models, covering routing, fallback, caching, rate limits, cost tracking and access control

Establish retrieval (RAG) as a platform service and engineer AI memory services. Cover ingestion, embeddings, hybrid search, reranking and access-controlled retrieval that cites its sources, ready for teams to plug in their own document sets. Handle context management, session and case memory, and long-term memory, with retention policies that meet data protection requirements

Develop the agent runtime (the harness). Cover orchestration, tool calling, MCP integrations, multi-agent workflows, sandboxed execution, approval checkpoints and audit trails

Embed guardrails and safety controls, making filtering, prompt-injection defences, personal data redaction and policy enforcement the default

Set prompt engineering, evaluation and AI development lifecycle (AI DLC) standards. Define prompt standards and supply tooling for golden datasets, automated evals, regression testing and red-teaming that teams can run against their own solutions. Establish automated, traceable CI/CD with evaluation gates, prompt and model versioning, controlled releases and drift monitoring, aligned with AI Programme governance and covering how engineers use AI coding assistants

Drive observability and operations. Implement tracing, dashboards and alerting across quality, latency, cost, safety and reliability,…