Machine Learning Engineer - National Security AI Start-Up - …, London

Machine Learning Engineer – National Security

Up to £98,000 | London or Gloucestershire (onsite) | SC/DV clearance

Few&Far is working with a specialist AI company building machine learning systems used in genuinely high‑stakes environments. This is applied ML in the real world - messy data, limited infrastructure, and problems where outcomes actually matter.

They’re looking for someone who enjoys taking ownership, working closely with end‑users, and building systems that hold up in production.

At a high level, you’ll be taking real operational problems and turning them into working machine learning systems that can be deployed and relied upon in the field.

That means:

• Working directly with users to understand problems and shape solutions
• Rapidly prototyping ideas, then turning them into production‑ready systems
• Deploying models in environments with constraints (compute, latency, connectivity)
• Writing clean, maintainable Python code that others can build on
• Contributing to technical decisions and, over time, supporting more junior engineers

They’re after someone who’s already spent time building and shipping ML systems, and is comfortable operating in environments where things aren’t always clearly defined.

In practice, that looks like:

• Confidence with Python and modern ML tooling (Docker, CI/CD, cloud, MLOps)
• A pragmatic mindset — able to balance speed, quality, and real‑world constraints
• Comfort working with non‑technical stakeholders and end‑users
• Eligibility for UK SC or DV clearance

Experience in secure or government environments is useful, but not essential. Similarly, exposure to areas like NLP, computer vision, LLMs, APIs, or data pipelines will help — but the core requirement is strong applied ML experience.

It’s a chance to work on problems that are genuinely meaningful, in a team that cares about building things properly — not just quickly. If you enjoy applied ML, ownership, and seeing your work used in the real world, this is a strong one to consider.