AI Engineer – LondonInvestment Management | AI & LLM Systems | Production OwnershipI’m representing a leading global outsourced investment office advising distinguished endowments, foundations, and prominent families, with approximately $60bn in assets under management. The firm is investing heavily in advanced AI capabilities to support both investment and operational teams.We’re hiring an AI Engineer to take end-to-end ownership of production AI features — from messy data ingestion to robust RAG pipelines and reliable APIs.The RoleYou’ll sit between “wrapper dev” and “researcher”: hands-on with Python and data, fluent in modern LLM tooling, and sceptical enough to design systems that assume models can fail.Key responsibilities include:Designing and maintaining RAG pipelines over unstructured data (PDFs, HTML, email, transcripts, APIs)Working with embeddings, vector databases (e.g. Pinecone, Weaviate, Qdrant) and graph databasesBuilding multi-agent / agentic workflows with strong conversation state managementIntegrating LLMs (OpenAI, Anthropic, open-source) with robust handling of rate limits, retries, context windows, and guardrailsImplementing prompt-injection defences, output filtering, and permissioned tool useDesigning evaluation frameworks (accuracy, faithfulness, latency) and regression testingShipping typed, well-tested backend services (Python/FastAPI; TypeScript a plus)Owning observability (logging, metrics, tracing, token usage) and production monitoringWhat We’re Looking For~3+ years software engineering experience, including 1–2 years building production AI/LLM systemsStrong Python (async, typing, testing, packaging, performance profiling)Hands-on RAG implementation (embeddings, semantic search, vector stores)Experience with at least one orchestration framework (e.g. LangChain, LlamaIndex)Solid grasp of LLM behaviour in practice (tokens, hallucinations, prompt injection risks)Experience designing evals and systematically reducing failure modesComfort with ETL/data preprocessing (pandas, SQL, HTML parsing)Cloud familiarity (Azure preferred; AWS/GCP welcome)Nice to have:Fine-tuning/adapting open-source modelsAI infrastructure and scalable model servingGraph databasesExperience in investment management or financial services