Director - AI Engineering

EMPLOYMENT TYPE:

Permanent DEPARTMENT:
Information Systems DIVISION:
Group WORKPLACE STRUCTURE:
Hybrid ROLE OVERVIEW PURPOSE OF JOBAs JTC enters the Genesis era of
growth and transformation under private equity ownership, the Group's ambition
is to become a data- and AI-powered specialist-services platform where every
interaction, acquisition and piece of expertise makes the firm smarter, more
efficient and easier to scale. The Director, AI Engineering leads the AI
Engineering & Data Science function within the unified Tech-Ops operating
model, the engineering engine room that turns JTC's AI & Agentic Strategy
2030 into working, governed and value-generating systems.

The postholder will build and
lead the team responsible for AI/ML model development and agentic engineering
across the three pillars of the AI strategy; AI for Service Delivery, AI for
Digital Products and AI for Productivity. This spans the JTC Agentic Platform
(MCP-enabled orchestration and validation across JTC systems and market tools),
the enterprise document extraction platform, end-to-end billing agents (Agentic
Thames and the Agentic Wheel), conversational AI analytics on the Klondike data
platform, function-specific JTC Agents and the Group-wide Agent Builder; all
delivered on a consolidated, Claude-based AI stack that preserves in-house
control and avoids vendor lock-in.

This is a hands-on engineering
leadership role. Acting as the Group's hub for AI in the Businesses and Group
functions, the postholder pairs deep technical credibility with commercial
judgement: every model and agent shipped must deliver measurable operational
ROI against the Technology Value Creation Plan, operate within the AI
Governance & Model Risk framework, and stay cost-disciplined through Fin Ops
controls. The ideal candidate combines frontier LLM and agentic engineering
expertise with the leadership skills to build and scale a high-performing team
in a PE-backed environment.

AI & Agentic Engineering Delivery Lead the design, development and production deployment of AI/ML models and agentic systems across the Group, delivering the prioritised use cases on JTC's two-year roadmap (55 use cases selected on value, feasibility and strategic fit) to enterprise quality.

Build and scale the near-term delivery priorities:
Agentic Thames (the end-to-end billing agent combining contract-reading, scoping, billing and timesheet quality-assurance AI skills), the Agentic Wheel billing transformation, AI-enabled tax workflow redesign, the standardised AI assistant, and AI-enabled analytics on the Klondike financial reporting Digital Integrated Agents for JTC's client-facing digital products, enabling non-linear user journeys, content and dynamic UI generation, conversational AI-enabled analytics and reporting, and persona-aware hyper-personalisation services.

Deliver the enterprise document extraction platform and agent
supporting workflows across service delivery, driving timeliness, accuracy and efficiency in middle- and back-office function-specific JTC Agents (e.g. Billing, Data Quality Control), individual AI productivity assistants with secure access to JTC data and systems and institutionalise the multi-platform Agent Builder for craft-your-own agents.

Agentic Platform & AI Architecture Own the engineering of the JTC Agentic Platform, MCP-enabled orchestration and validation across JTC systems and market tools, maintaining in-house control of the agentic layer to avoid LLM lock-in.

Build around the consolidated Claude stack (Claude, Cowork, Claude Code and the developer platform), migrating existing agentic workloads onto it and prioritising app-agnostic agents on the JTC Agentic Layer over vendor-embedded AI.Partner with Data Engineering & Provisioning to ensure AI-ready data, reusable, governed structured and unstructured data pipelines feeding models, agents and analytics across the Group.

Define
and enforce AI engineering standards, deployment patterns, integration patterns, testing and evaluation frameworks, and contribute AI/data positions to the Architecture Forum's buy-vs-build and standards decisions.

Evaluate third-party embedded AI and SaaS agents (e.g. Safebooks.ai, Rossum) for benefit and cross-application scalability, integrating those that earn their place on the & Value Delivery Co-own use-case prioritisation with the AI Product Management function, impact and feasibility assessment, quarterly roadmap refresh, dependency and sequencing management, and capacity and…