AI Engineer, Forensic and Financial Crime Technology Advisory and Data Analytics

Overview

In this role you design and deliver end-to-end GenAI-enabled solutions for financial crime processes within Deloitte’s advisory practice. You will lead architecture decisions, craft agentic workflows, and build production-grade data pipelines, collaborating with cross-disciplinary teams. The work centers on transforming complex financial crime challenges into reliable AI-driven outcomes in regulated financial services. You’ll contribute to a growth-focused, innovative culture and develop your expertise through active client-facing and mentorship opportunities.

Pay / Benefits

• hybrid working in London
• flexible working arrangements
• professional development
• wellbeing support
• opportunity to work with leading clients
• collaborative culture

Responsibilities

• Define and integrate solution modules (frontend, backend, APIs, data stores, caching, memory, logging) and determine AI placement within the architecture
• Design and evaluate agentic workflows and prompts, ensuring reliable, production-ready behavior
• Build and own code and data pipelines, including data ingestion, parsing, chunking and embedding
• Create architecture artefacts (C4, sequence diagrams) and choose data stores (relational, NoSQL, vector) and messaging strategies
• Lead end-to-end AI/GenAI projects, including architecture, coding, testing, and deployment
• Measure accuracy and establish evaluation regimes (golden datasets, regression suites, tracing)
• Collaborate with cross-functional teams and mentor others to deliver client-ready solutions
• Utilize a range of tools for design, development, and evaluation (see external_tools)

Key requirements

• Proven experience designing and shipping software or AI solutions end to end
• Strong system-design skills with ability to decompose problems across frontend, backend, APIs, data, caching, memory, async/eventing, logging
• Hands-on experience building LLM/GenAI applications, including RAG, tool use, orchestration, and multi-agent patterns
• Practical prompt engineering for production, using structured outputs and tool calling
• Experience evaluating non-deterministic systems with golden datasets, regression testing and observability tracing
• Strong coding ability in Python and/or TypeScript, with Git/PR workflows and automated testing
• Data engineering fundamentals: relational, NoSQL and vector stores and document/ETL pipelines
• Fluent use of AI coding assistants while maintaining code ownership and testing described
• Ability to lead workstreams and communicate complex designs to technical and non-technical audiences

• Leadership and mentoring
• Strong communication with senior clients
• Collaborative teamwork
• LangChain/LangGraph
• LlamaIndex
• Ragas

…