Agentic systems
Design and ship tool-using agents, multi-agent workflows, MCP integrations, memory, evaluation, observability, and recoverable fallbacks.
Contract Lead AI Engineer · London / UK remote
Production AI needs a lead who can ship.
I join product and engineering teams as a hands-on Contract Lead AI Engineer — turning agentic and LLM use cases into secure, observable systems that run in production.
- 15+
- years delivering software
- 4
- years focused on agentic AI
- 3
- live agentic platforms
- 115
- roles in a multi-agent workspace
What I deliver
Architecture, code, and adoption in one contract.
The useful middle between a strategy consultancy and a narrow implementation resource: senior enough to shape the system, hands-on enough to build it.
Enterprise AI integration
Connect LLM applications to real data, APIs, controls, and business workflows — with security, audit, and human review designed in.
AI-enabled delivery
Improve the full engineering loop with Claude Code and coding agents: context, planning, implementation, review, testing, and deployment.
Selected enterprise delivery
Regulated systems. Cross-functional teams. Production outcomes.
Allianz
AI Consultant · Technical PM
Delivered Copilot Studio assistants, voice agents, and AI-driven contact-centre workflows in a regulated environment, including GDPR, redaction, and audit requirements.
RWS
Technical PM · AI Consultant
Integrated AI translation workflows into an established language-services platform and led delivery across cross-region engineering teams.
Sainsbury’s
Technical Delivery Manager
Delivered a CI/CD data pipeline using Snowflake, Jenkins, and Airflow to support marketing analytics at enterprise scale.
First 30 days
Get one system through the whole loop.
- 01Choose the production outcome
Map the workflow, data, users, controls, and success measure. Cut the programme to one vertical slice that can prove value.
- 02Ship the thin end-to-end slice
Build through the real stack — model, tools, API, data, evaluation, observability, review gate, and deployment path.
- 03Scale what survives contact
Use evidence from real runs to improve quality, latency, cost, and control. Leave the team with code, tests, decisions, and a runbook.
Built for accountable delivery
- Human release gates for consequential actions
- Evaluation and acceptance evidence
- Cost, latency, and failure visibility
- Code, tests, decisions, and runbook handed over
What must be live in 90 days?
Send the outcome, current stack, team shape, location, duration, and IR35 status. I reply within one working day.