Deployed AI Engineer

The Role

This role is the technical spearhead of our consultancy client's new-world work. You'll operate much like a startup CTO, working in a small pod, typically alongside a Delivery Lead, and taking ownership of the end-to-end technical execution of client engagements, from scoping and system design through to build and rollout.

This is genuinely forward-deployed work. You'll be embedded within the client's environment, working with their real data and operating within their security and technology landscape.

The expectation is to have a working prototype within the first few days of an engagement, with feedback and iteration happening every couple of days. Successful POCs are then taken through to MVP within weeks.

Between engagements, you'll help extend our internal AI platform and turn what you've proven in the field into accelerators, templates and playbooks - making each engagement faster, smarter and more effective than the last.

What You'll Be Doing

• Building working AI proofs on real client data during Assess engagements - the prototype is the discovery tool, built while the value case is made
• Delivering client POCs and MVPs: RAG pipelines, agent architectures, LLM integrations and protocol-driven tooling (MCP, tool orchestration)
• Iterating in short loops: demoing every few days, taking feedback, changing course without ceremony
• Engineering for production from day one: guardrails, security and integration into the client's estate, scalability and telemetry - designed inside the build, not bolted on
• Proving trustworthiness with evals: golden datasets, automated evaluation pipelines, accuracy and drift monitoring - and hill-climbing on the results
• Wrangling client data: pipelines, messy edge cases, integrations that are harder than they look
• Owning and evolving our internal AI platform; codifying repeatable field patterns into reusable assets.
• Leading advanced technical sessions in our upskilling programme, and upskilling client engineers during Transform engagements
• Evaluating emerging AI tools, frameworks and protocols, and making pragmatic adoption calls
• Strong software engineering foundations - clean code, testing, CI/CD, production mindset - with strong general-purpose programming (e.g. Python, TypeScript) and proficiency in 2+ modern languages
• Full-stack capability: enough front-end, back-end and data engineering to build the whole thing yourself
• Hands-on production LLM and agent experience: prompt engineering, agent workflows, RAG, tool orchestration and MCP - with evidence you've shipped AI users actually rely on, and how you proved it could be trusted
• Evaluation-driven habits: golden datasets, eval frameworks, guardrails - you measure whether it works rather than asserting it does
• Client-facing delivery experience, and comfort with the constraints of enterprise environments: security, compliance, legacy integration
• High agency and comfort with ambiguity - you can operate with minimal supervision in a client's world
• A value instinct: you can explain what a build is worth in business terms, and say when something isn't worth building
• Vector databases, embeddings, fine-tuning
• Former founder or startup experience
• Open-source contributions or visible AI side projects

#J-18808-Ljbffr

Full job description from WhatJobs. Apply for this role →