AI Engineer - Legal Innovation

About the role

Norton Rose Fulbright is building a new R&D capability focused on data-driven and AI-enabled solutions for legal services. The AI Engineer builds end-to-end AI capabilities for R&D products such as agentic workflows, retrieval and context, agent and evaluation harnesses and their user-facing features. This is a hands-on full-stack AI engineering role within a small, multidisciplinary team.

Key responsibilities

• Build complete AI-enabled product features across UI, backend, AI orchestration, retrieval and approved integrations.
• Design agentic AI workflows with tool calling, document processing and human-review steps.
• Build reusable agent harnesses covering tools, state, context, approvals, traces and test fixtures.
• Build grounded retrieval and context patterns with search indexes and document repositories, providing evidence and citations.
• Design evaluation harnesses with test sets, automated and human grading, regression tests and user feedback.
• Manage prompts, model configuration, tool schemas and routing, including fallbacks and cost/latency trade-offs.
• Implement observability via traces, logs, metrics and evaluation results.
• Build APIs and user interfaces that show AI output, confidence, evidence and next actions.
• Write maintainable Python and TypeScript code using Docker, Git, automated testing and CI/CD, applying security and deployment patterns.

Required profile

• At least three years' experience delivering production AI products, with recent SaaS experience.
• Demonstrable experience with LLM APIs, structured outputs, retrieval, embeddings, tool calling and agents.
• Experience evaluating AI workflows and moving features from prototype to production.
• Experience with API design, authentication, authorisation and secrets handling.
• Experience with AI observability, tracing and evaluation tooling.
• Experience in legal technology or a law firm is not required but advantageous.

Required skills

• Python and TypeScript or JavaScript; FastAPI, Flask or similar.
• React, Next.js or equivalent for interfaces.
• Docker, Git, automated testing and CI/CD.