AI Engineer in London

Location: Manchester / Hybrid / Remote - depending on candidate location. If you are not based in Manchester, you will need to travel to Manchester once or twice a month.

Our global client is building advanced behavioural intelligence technology that enables secure, adaptive digital identity. By analysing how people naturally interact with devices, their AI systems create powerful authentication signals designed for real-world use at scale. This is a high-impact opportunity to join a significantly growing AI team and take ownership of creating and working on cutting-edge models and pipelines.

The Role

As an AI Engineer, you will design, build, and refine machine learning models that sit at the heart of the company's behavioural AI platform. This is a hands-on role involving working with real sensor data from devices we use daily, developing and training architectures, and deploying models that make authentication decisions in production. You'll collaborate closely with other AI engineers, as well as engineering and product teams, to ensure models are robust, efficient, and production ready.

Key Responsibilities

• Develop and train deep learning models for behavioural authentication, working with multimodal sensor data including accelerometer, gyroscope, touch patterns, and device interaction signals.
• Build data processing pipelines for irregular, event-driven time series data from mobile devices.
• Design and run experiments to improve model performance on key authentication metrics (False Accept Rate, False Reject Rate).
• Contribute to the evolution of large-scale behavioural modelling approaches and shared training systems.
• Prepare models for efficient on-device execution, balancing performance with mobile hardware limitations.
• Deploy models for edge inference using CoreML and ONNX, optimising for mobile device constraints.
• Work closely with mobile engineering teams to embed AI functionality into production SDKs.
• Define and improve methods for evaluating, benchmarking, and validating behavioural authentication systems.
• Contribute to Large Behavioural Model architecture and training infrastructure.

What We're Looking For

Required

• Strong experience building deep learning systems using PyTorch - not just using API wrappers or pre-trained models.
• Experience implementing modern neural architectures, including attention-based models, custom model heads and positional encoding.
• Experience working with temporal or sequential data, particularly from sensors or user interactions, wearables etc.
• Comfortable managing experiments, model versions, and reproducible ML workflows.
• Experience deploying machine learning models using cloud infrastructure (AWS preferred).
• Strong Python skills and a practical, delivery-focused mindset.

Desirable

• PhD in Machine Learning, Computer Science, Applied Mathematics, Computational Linguistics, or a related field.
• Experience with behavioural modelling, biometrics, authentication systems, security applications or security-focused AI.
• Experience with behavioural data, human activity recognition, or gait analysis.
• Exposure to deploying models on-device or in constrained environments.
• Familiarity with representation learning or self-supervised approaches.
• Research background or publications in relevant areas.

Tech stack

• ML/AI: PyTorch, MLflow, SageMaker, ZenML.
• Infrastructure: AWS, Kubernetes, Docker.
• Edge deployment: CoreML, ONNX.
• Data: Python, S3, multimodal sensor data pipelines.
• Collaboration: JIRA, Git, structured OKR methodology.

Why you will enjoy working with our client:

You'll join a small, growing AI team where engineers have real ownership and autonomy. You'll be trusted to tackle complex, open-ended problems, collaborate closely across disciplines, and apply research thinking to systems that are built to ship. It's an environment that values curiosity, delivery, and continuous learning.