Defence AI Engineer - Production LLMs & Agentic Systems

Understanding Recruitment Ltd; Greater London; Hybrid; GBP 85,000 - 120,000; Full time

Job summary

Understanding Recruitment Ltd in the United Kingdom is seeking an AI Engineer to build and deploy production-grade agentic systems for defence and national security environments, focusing on performance, reliability and real-world impact.

The role involves working with LLM orchestration, model optimisation, multimodal AI, real-time sensor data, and ownership from research to production. Strong ML foundations and production experience are essential.

Qualifications

• Strong foundations in machine learning, deep learning and applied mathematics.
• Production experience building agentic AI / LLM-based systems.
• Strong programming ability in Python, Rust, C++ or Java.
• Experience with PyTorch, JAX, TensorFlow or agent orchestration frameworks.
• Excellent engineering judgement and the ability to turn research into robust, production-ready AI.
• Interest in solving difficult technical problems with real-world consequences.
• PhDs, top-tier ML publications and experience deploying AI in constrained or regulated environments are a bonus.

Responsibilities

• Design and deploy production-grade agentic AI systems.
• Develop LLM orchestration, agent architectures and model optimisation.
• Work with multimodal AI, computer vision and real-time sensor data.
• Take ownership from research and prototyping through to production.
• Ensure reliability, safety, scalability and deployment of AI systems.

Skills

Machine learning

Deep learning

Applied mathematics

Production-grade AI

Problem solving

Tools

Python

Rust

C++

Java

PyTorch

JAX

TensorFlow

Agent orchestration

Job description

Understanding Recruitment Ltd in the United Kingdom is seeking an AI Engineer to build and deploy production-grade agentic systems for defence and national security environments, focusing on performance, reliability and real-world impact.

The role involves working with LLM orchestration, model optimisation, multimodal AI, real-time sensor data, and ownership from research to production. Strong ML foundations and production experience are essential.