Applied ML Technical Lead, Defence

Job title: Applied ML Technical Lead, Defence

Location: London; full in-office working as default

Start date: ASAP

Reports to: CEO

Compensation: Salary and Equity options available for discussion

Clearance: Must be eligible to hold UK Security Vetting Clearance SC or above (no prior clearance required)

Application Deadline: 08/11/26 at 23:59 UK Time

___________________________________________________________________________

Help put frontier AI to work in Defence

Cosine is a UK AI lab building models, applications and training infrastructure for mission-critical organisations.

We build coding agents that work across real software systems, specialised Lumen models, AI-native security products and the infrastructure needed to deploy advanced AI in demanding environments. Our customers include some of the world's largest enterprises, including organisations operating behind strict security, regulatory and technical constraints.

We're also training Lumen Sovereign, the first frontier large language model trained in the UK, with the support of the UK Government.

We're now looking for an Applied ML Technical Lead to help develop and deploy advanced AI capabilities for organisations operating in UK Defence and national security.

This is a hands-on technical leadership role. You will work across machine learning development, deployment and customer-facing technical delivery. You may spend part of your time training and evaluating models on complex sensor data, and part of your time working with Defence stakeholders to understand operational needs, shape priorities and guide delivery.

You will not be expected to act as a traditional project manager. Instead, you will own the technical direction and day-to-day progress of applied ML work, coordinating the people and decisions needed to turn promising research into useful operational capability.

___________________________________________________________________________

What you'll do

Build ML capabilities for real Defence problems

Design, build, train and evaluate machine learning models for challenging applied Defence and national security use cases.

Work on problems involving complex sensor data, signal processing, pattern recognition, classification and multimodal intelligence.

Develop and improve model architectures, training pipelines, evaluation frameworks and inference systems.

Work with real-world and synthetic datasets to improve model accuracy, robustness and generalisation.

Translate research prototypes into reliable, deployable capabilities that operate within demanding technical and security constraints.

Work closely with Cosine's ML research and engineering teams while taking direct ownership of technical outcomes.

Lead technical delivery from prototype to deployment

Own the technical direction and day-to-day progress of applied ML projects.

Turn operational objectives into concrete technical goals, experiments, milestones and success measures.

Coordinate work across ML researchers, software engineers, infrastructure teams and external stakeholders.

Identify technical dependencies, delivery risks and blockers, and help drive them to resolution.

Establish clear evaluation criteria and demonstrate progress through measurable improvements and working capabilities.

Balance hands-on engineering with the planning and coordination needed to keep complex technical work moving.

Work directly with Defence customers

Engage with technical and operational stakeholders across the UK Defence and national security ecosystem.

Understand operational challenges and identify where machine learning can make a meaningful difference.

Translate customer requirements and constraints into technical objectives, datasets, experiments and delivery plans.

Present technical progress, demonstrate working capabilities and communicate results clearly to technical and non-technical audiences.

Build trusted relationships with Defence stakeholders and act as a credible technical counterpart.

Help customers understand the practical considerations of integrating and deploying ML systems in their existing environments.

Turn research into operational capability

Help deploy ML models into real-world environments, including secure, on-premise and resource-constrained systems.

Work on model optimisation, inference performance, integration and reliability.

Develop robust approaches to evaluation, validation and continuous model improvement.

Understand data availability, access requirements, security restrictions and deployment constraints.

Establish feedback loops between operational users and Cosine's engineering teams.

Ensure that what we build is not only technically impressive, but genuinely useful to the people relying on it.

___________________________________________________________________________

What we’re looking for

You'll probably have a background in applied machine learning, ML engineering, research engineering or a similarly technical role, along with experience taking ownership of complex technical work.

You should be comfortable:

Writing production-quality Python and working with modern ML frameworks such as PyTorch.

Building, training, fine-tuning and evaluating machine learning models.

Working with complex datasets, designing experiments and interpreting model performance.

Understanding the trade-offs between model performance, compute requirements, latency and deployment constraints.

Working with engineering infrastructure, including Git, containers, Linux and GPU compute.

Taking technical ownership of projects with multiple stakeholders, uncertain requirements and evolving priorities.

Communicating technical concepts clearly to researchers, engineers, senior stakeholders and operational users.

Moving between model development, customer discussions and delivery planning without losing technical depth.

Making sound technical judgements, prioritising effectively and finding practical solutions wh