Key responsibilities

• Build agent workflows and RAG pipelines over finance policies, contracts, trade agreements and regulatory text

• Implement tool-calling into ERP, data platform and reporting APIs, with approval gates for write actions

• Design, version and test prompts, context and memory strategies like code

• Build evaluation suites (accuracy, grounding, safety, regression) and wire them into CI

• Implement guardrails: PII redaction, prompt-injection defence, grounding and citation checks

About you

• 5+ years of software engineering with strong Python, including 1+ year on LLM / agent systems in production

• One or more agent frameworks (LangGraph, Semantic Kernel, AutoGen, CrewAI or similar)

• RAG with a vector or hybrid search store (e.g. Azure AI Search, pgvector, OpenSearch)

• Evaluation tooling (e.g. Ragas, DeepEval, promptfoo), API / cloud fundamentals, Git and CI/CD

• Domain: CPG / FMCG and Finance Transformation