UKI Finance RP

Finance AI Architect

The practice

Finance is one of the most demanding and valuable environments in which to apply modern technology. You will work with complex enterprise data, mission-critical processes and high-impact decisions, using AI, data and engineering to reshape how organisations plan, control performance and allocate resources. The opportunity goes beyond building technically strong solutions: you will see how those solutions influence cash, profitability, risk and business growth, and take them from experimentation into trusted, production-ready capabilities. Working in Finance Reinvention allows you to remain close to leading-edge technology while developing an understanding of the CFO agenda, gaining exposure to senior decision-makers and building the commercial judgement needed to solve enterprise-wide challenges. This combination of deep technical capability, finance-domain expertise and measurable business impact creates a differentiated career path that is difficult to develop in a purely technology-focused role.

Purpose of the role

Technical design authority for agentic finance solutions across the delivery portfolio. Accountable for the end-to-end architecture from early shaping through production transition, including agent topology, orchestration patterns, runtime selection, guardrails and cost to serve, together with retrieval and memory, enterprise integration, security, resilience, observability and human oversight. The role operates across several pods concurrently rather than being embedded in a single delivery team, and provides technical credibility with client CTOs and heads of architecture, as well as CIOs, Chief Data and AI Officers, and CFO technology leaders.

Responsibilities

• Define and maintain the reference architecture for finance agents, covering orchestration, tool and MCP design, memory, retrieval, human handoff, evaluation and rollback, including RAG and graph-based retrieval, context engineering, model gateways, lifecycle versioning and graceful degradation.

• Set model selection and routing strategy across proprietary and open-weight models and cloud AI platforms, with a defensible position on cost, latency and reliability, as well as quality, safety, data residency, vendor lock-in and supportability.

• Design the control plane, including segregation of duties, approval gates, evidence trails and auditability, as a first-class part of the architecture, with identity and access controls, policy enforcement, prompt and model traceability, sensitive-data handling and explainability.

• Own the non-functional architecture and technical assurance for scalability, resilience, security, privacy, accessibility, latency, throughput, disaster recovery and cost efficiency.

• Define LLMOps and AgentOps standards and production acceptance criteria, including evaluation suites, red teaming, monitoring, service levels, feedback loops, release management, incident response and rollback runbooks.

• Define secure integration patterns for ERP and EPM platforms, data and content services, workflow tools, APIs and event-driven systems, working closely with client security, data and enterprise architecture teams.

• Provide technical authority in pursuits and lead the technical response to CFO and CIO stakeholders from discovery and proof of value through production, including architecture decisions, technical risks, estimates, dependencies and transition plans.

• Establish the engineering standards that AI Engineers build to, and develop technical depth across the team through reusable patterns, architecture decision records, design reviews, exception governance and mentoring.

• Define measurable solution outcomes with finance and value leads, including quality, automation and exception rates, control effectiveness, user adoption and cost to serve.

Essential experience

• Production agentic or LLM systems delivered in an enterprise setting, with demonstrable evidence of operation at scale and of issues encountered and resolved, including moving solutions from proof of concept into controlled production and responding to reliability, security or safety incidents.

• Substantial architecture background covering distributed systems, integration, API and event design, and cloud platform experience on Azure, AWS or GCP, DataBricks, Snowflake, Palantir etc. , including identity, secrets, encryption, deployment, observability and production operations.

• Hands-on depth in LLM application architecture, including RAG, embeddings and vector or graph retrieval, tool and function calling, structured outputs, orchestration frameworks and model gateways; sufficient engineering fluency to prototype or challenge implementation choices.

• A developed position on evaluation and reliability engineering for LLM systems, including offline and online evaluation, groundedness and task-quality measures, failure-mode analysis, monitoring and cost or latency optimisation.

• Working knowledge of responsible AI and secure AI delivery, including threat modelling, privacy and PII controls, access control, auditability, model risk, GDPR and EU AI Act considerations.

• Credibility with senior client stakeholders under technical challenge, with the ability to explain trade-offs to both technical leaders and finance executives and to lead technical shaping in pursuits.

• Ability to translate finance-process, control and audit requirements into executable architecture and engineering acceptance criteria.

• At least 6 years relevant professional experience

Desirable

• Enterprise finance exposure across ERP, financial close, controls or audit, EPM, planning or finance transformation.

• Delivery experience in regulated environments, including EU AI Act or equivalent obligations, model-risk management or external assurance.

• Experience with semantic or knowledge-graph architectures, GraphRAG, data-product patterns or managed AI services.

• Relevant cloud, architecture, security or AI platform certifications.