GenAI & Agentic System Engineer

GenAI & Agentic System Engineer- 1 year Contract What You'll Do Design and build Retrieval-Augmented Generation (RAG) pipelines - chunking, embeddings, semantic search, and retrieval tuning for accuracy and grounding. Develop AI agents and multi-step workflows using LangChain / LangGraph, including tool calling and orchestration across multiple steps. Build and maintain backend services and REST APIs (FastAPI / Flask) that expose AI capabilities to internal and member-facing applications. Integrate and optimise usage of LLM providers (OpenAI, Anthropic, Gemini), balancing quality, latency, and cost. Work with vector databases (Pinecone, Weaviate, Chroma, FAISS, or Qdrant) and relational/document stores (PostgreSQL / MongoDB) for storage and retrieval. Apply strong prompt engineering practices and iterate based on evaluation results. Containerise and deploy services with Docker, and contribute to Git-based CI/CD workflows. Set up evaluation and quality checks for AI outputs, and continuously improve reliability and accuracy. GenAI & Agentic System Engineer- 1 year Contract What You'll Do Design and build Retrieval-Augmented Generation (RAG) pipelines - chunking, embeddings, semantic search, and retrieval tuning for accuracy and grounding. Develop AI agents and multi-step workflows using LangChain / LangGraph, including tool calling and orchestration across multiple steps. Build and maintain backend services and REST APIs (FastAPI / Flask) that expose AI capabilities to internal and member-facing applications. Integrate and optimise usage of LLM providers (OpenAI, Anthropic, Gemini), balancing quality, latency, and cost. Work with vector databases (Pinecone, Weaviate, Chroma, FAISS, or Qdrant) and relational/document stores (PostgreSQL / MongoDB) for storage and retrieval. Apply strong prompt engineering practices and iterate based on evaluation results. Containerise and deploy services with Docker, and contribute to Git-based CI/CD workflows. Set up evaluation and quality checks for AI outputs, and continuously improve reliability and accuracy.