LLM Engineer

Context

Our customer is standing up a dedicated AI team to bring structured AI engineering capability to its software delivery organisation. The team is split into two functions: a Platform team that builds shared infrastructure, and an Enablement team that works with engineering and product teams to ship AI use cases and transfer the capability to do so independently. The goal is measurable AI capability across the SDLC within six months.

This role sits in the Enablement team and builds the patterns, harnesses, and reusable components that underpin AI use cases across the SDLC. The work happens inside real use cases alongside delivery teams. Artefacts produced here feed into a shared library that teams draw on as they build independently.

Missions

• Design and implement RAG pipelines: retrieval strategies, chunking, re-ranking, hybrid search. Contribute to storage and vector layer decisions.

• Develop agent architecture: tool use patterns, multi-step reasoning flows, human-in-the-loop design, context window management.

• Build evaluation harnesses: test scaffolding created in coordination with the LLMOps engineer.

• Perform prompt engineering: structured prompting, few-shot design, chain-of-thought patterns.

• Create and maintain a shared component library: reusable TypeScript components emerging from use cases that feed into the Salomon AI SDK alongside the platform team's developer surface.

• Participate in use case kickoffs and builds alongside the FDE engineer and delivery team contributors.

Tools & Environment

• Primary stack: TypeScript and Node.js

• Representative environments: Vercel AI SDK, AWS Bedrock SDK

• Primary model: Claude via Anthropic API and Claude Enterprise

Working Conditions

• Collaboration with product managers, domain engineers, and platform contributors to build alongside people with different specialisms.

• Delivery of four to six use cases to production within six months with documented quality baselines.

• Codification and reuse of core patterns.

• Enablement of engineering and product teams to deliver AI work independently after engagement.

1. Experience with LLM systems in production, taking LLM-powered systems from prototype through evaluation harness to production deployment
2. RAG and retrieval pipeline design experience, including understanding trade-offs between dense, sparse, and hybrid approaches
3. Experience with agentic patterns involving tool-calling and multi-step agent flows, including reliability characteristics
4. Ability to design meaningful evaluation datasets for unfamiliar domains and write evals as part of the build process
5. Proficiency in TypeScript and Node.js, with familiarity with Vercel AI SDK and AWS Bedrock SDK environments
6. Collaboration experience working alongside product managers, domain engineers, and platform contributors