AI / LLM engineer

Role overview

An engineering role focused on building AI and LLM-powered systems that stay grounded in real client data. The work spans copilots, retrieval-augmented generation pipelines, and the evaluation loops that keep those systems honest over time. A strong emphasis is placed on guardrails, cost control, and thoughtful human review where it matters most.

Responsibilities

- Design and build LLM-powered features such as copilots, assistants, and retrieval-augmented generation pipelines. - Develop evaluation loops that measure quality, grounding, and safety across model iterations. - Implement guardrails that constrain model behavior, protect sensitive data, and bound costs. - Integrate human review checkpoints where outputs carry meaningful business or user risk. - Collaborate with data and product teams to keep systems anchored in client-provided data sources.

Requirements

- Practical experience building production systems that use large language models. - Familiarity with retrieval-augmented generation patterns, embeddings, and vector search. - Comfort designing evaluation harnesses for model behavior, regression detection, and quality assurance. - Awareness of cost dynamics in LLM systems, including token usage, caching, and model selection. - Pragmatic mindset around safety, grounding, and where human review genuinely adds value.

Nice to have

- Experience with prompt engineering, fine-tuning, or model selection across providers. - Exposure to AI observability, tracing, and structured output validation frameworks.