The work

• Build functional end-to-end agents with Python, LLM frameworks, APIs and databases.
• Set up vector/relational databases and containers for real-world testing; choose RAG, function-calling or API-wrapper architectures.
• Iterate on business feedback and discard unviable concepts.
• Build evaluation pipelines for accuracy, hallucination rates, latency and cost.
• Create ground-truth labelled datasets with business owners; analyse differences between agent outputs and human pilot decisions.
• Assess operational workflows and new AI architectures with strategy teams.
• Translate validated MVPs into technical specifications, architecture diagrams and data flows for core Technology.
• Partner with the AI platform team to absorb MVP logic into internal SDKs.

Hands-on engineering in the AI Optimization team of a flight-technology business. The role owns prototype construction and evaluation; the core Technology team is responsible for scaling validated work. No people-management remit is documented.

What you bring

• Must have:3+years of software engineering with strong Python and rapid-prototyping skills.
• Must have:hands-on LLM orchestration such as tool use, RAG or structured outputs; named frameworks/APIs are examples.
• Must have:API design, SQL, relational/vector databases and evaluation-framework experience.
• Must have:pragmatic fast experimentation and ability to turn complex business logic into concise technical specifications.
• Nice to have:FDE/solutions-architecture background, modern deployment stacks, travel/airline/payments knowledge and basic product-management/ROI/user-validation understanding.