Forward Deployed Engineer (Production LLM Systems)

? London, hybrid

? Permanent

About the Client

Our client is a well established UK business operating in a regulated sector, with a large customer base and a strong engineering culture. They are investing seriously in AI and are building a dedicated engineering team to take it from experiments into systems the business relies on every day.

 

The Role

 

Teams across the business have already started building their own AI tools, and this hire will give that work a safe route into production. You will work across three areas:

• Partnering with business teams to automate and improve how they work
• Taking models built by the in house ML team into live customer features
• Turning what you learn into shared tooling the whole company can use

The focus is building dependable software that happens to use LLMs, so you will spend far more time on engineering than on prompts or model training.

 

What You Will Do

• Work directly with business teams to understand their problems, define the right solution, then design, build and deploy it
• Take projects from first prototype through to stable production, and agree clear ownership once they are live
• Build Python pipelines that combine deterministic steps with LLM calls, using good judgement on where each one belongs
• Put evaluation, guardrails, monitoring and cost controls around everything you ship
• Rebuild useful tools that teams have created for themselves into properly owned business systems when they prove their worth
• Deploy models built by the data science team into customer facing production, covering serving, integration and everything around the model
• Spot the patterns across your projects and turn them into shared tools and safe defaults that other teams can pick up and use themselves
• Learn from specialist engineers during the early stages of the team and help keep that knowledge in house

What Success Looks Like

 

Early on, you will pick your first projects with the business, ship them, and be able to point to clear results. You will also help get one of the data science team's models running reliably for customers.

What You Will Bring

 

• Experience building for people outside engineering, whether as a consultant, a forward deployed engineer or on an internal platform team
• Comfort taking a vague problem and turning it into a scoped, deliverable system with minimal direction
• 5+ years of software engineering experience with real ownership of production systems
• A track record of building and shipping LLM based systems that have run in production
• Strong production Python as your main language
• Hands on experience working with models through APIs, including orchestration, structured output, tool use, agents and RAG where it fits
• Experience designing evaluation sets, measuring the accuracy of LLM steps, and regression testing prompts and workflows
• Practical safety experience, including PII handling, keeping data within the right boundaries, prompt injection awareness and human in the loop design
• Experience keeping LLM systems fast and affordable, whether through smarter model selection, caching or batching
• CI/CD, containerisation, and monitoring and alerting for AI workloads
• Solid data processing fundamentals, including pipelines, transformation, validation and handling anomalies