AI Engineer (Agentic Systems)

We are looking for an AI Engineer to join our core team to accelerate the development of the Governa AI product.

This is not a role for someone who simply integrates API endpoints. We are building complex, multi-agent systems that require a deep understanding of state management, concurrency, reliability, and production-grade AI architecture.

You will be responsible for architecting agentic workflows, optimising our Retrieval-Augmented Generation (RAG) pipelines, and ensuring that AI interactions are performant, secure, and observable in production environments.

Technical Stack & Responsibilities

1. Agentic Design & Orchestration

• Design and implement sophisticated agentic workflows using LangChain and LangGraph
• Apply agentic design patterns (Reflection, Planning, Tool Use) to solve non-linear problems
• Build and maintain high-performance backend services using FastAPI
• Optimise vector databases and embedding strategies to improve retrieval precision
• Engineer complex SQL queries with strong emphasis on efficiency and safety (including SQL injection prevention in LLM-driven contexts)
• Manage data pipelines and PII handling to maintain strict privacy and compliance standards

3. Performance & Reliability

• Implement concurrency patterns to handle high-throughput LLM interactions
• Utilise observability and monitoring tools to trace agent reasoning and debug hallucinations
• Refine prompt engineering strategies and LLM integrations (OpenAI and open-source models) to ensure consistent, high-quality output

What We’re Looking For

• 2-3 years of professional experience specifically in AI/ML engineering, or advanced backend roles with strong LLM exposure
• Advanced Python mastery – including async programming, type hinting, and performance profiling
• Architectural mindset – you design scalable systems, understand HTTP/REST best practices, and can scale distributed services
• The 'Agentic' edge – direct experience with LangGraph or similar state-machine-based orchestration frameworks is highly preferred
• Security-first thinking – knowledge of PII scrubbing and safe data handling within AI-driven systems