Artificial Intelligence Jobs

• Exeter
• £21,805 pa
• Contract

Investigates how two neuroscience developments—astrocyte-neuron interactions and spontaneous retinal waves—can improve deep networks. The work involves adding astrocyte-inspired regularisation to training and using synaptic plasticity rules for network initialisation, then testing generalisation, robustness, explainability, learning speed and data efficiency.

• Penryn
• On-site
• £21,805 pa
• Entry level
• Contract

Develops physics-aware spatio-temporal graph neural networks for 0–6-hour visibility forecasting, focusing on low-visibility events in southeast England. The work spans six packages, from benchmarking NWP and ML baselines to linking learned representations with atmospheric physics, testing robustness to degraded observations, and operational evaluation with Met Office scientists.

STGNNs · NWP

• Ireland
• Remote
• Lead · manages people

Leads a team of three data scientists building machine learning and agentic AI systems for threat detection, malware analysis and security automation. Sets technical standards and strategy for ML development, deployment and monitoring, while staying hands-on with Python, big data and cloud infrastructure.

AWS · Linux

• Ireland
• Remote
• Senior · manages people

Lead a team of data scientists working at the intersection of machine learning, large language models and cybersecurity, translating business needs into technical objectives and driving research into production ML systems. The role combines people leadership, roadmap planning and stakeholder communication with hands-on work on malware detection and threat analysis at massive scale.

AWS · Linux

• London
• Hybrid · 1 day in office
• £46,000 – £52,000 pa
• Contract

Join a national security team building robust AI agents and realistic training environments to defend critical infrastructure from cyber threats. Day-to-day work spans realistic network simulation environments, novel agent architectures, and evaluation tasks, requiring strong Python, Linux and git skills plus a research mindset. Hybrid role with just one day per week in the central London office.

Python · Git · Linux

• Cambridge
• On-site
• £45,000 – £75,000 pa

Design electro-optical camera systems including lenses, mirrors and sensors, optimising them for VIS, NIR and FIR applications. Work spans simulations in OpticStudio, component selection, prototype device testing and process documentation, taking projects from concept through to product release. The role is fully onsite and subject to security clearance due to military applications.

OpticStudio

• London
• Hybrid
• Lead
• Permanent

Lead the technical direction for an offline 3D reconstruction system that turns raw vehicle sensor data into quality-assessed 3D ground-truth geometry and ego motion. The role spans SLAM, lidar point-cloud registration, state estimation and sensor fusion, with responsibility for architecture, roadmaps, prototypes and cross-functional alignment across perception, calibration and validation teams.

ROS · PCL · OpenCV +3

• London
• Permanent

Conducts research on safety post-training methods such as RLHF, RLAIF, DPO, PPO, reward modeling, adversarial training and synthetic data, then builds evaluations for harms like jailbreaks, hallucination and policy adherence. Works cross-functionally with core training, product, policy, legal and trust and safety teams to ship safety improvements into production models.

OpenAI

• Filton
• £45,000 – £58,000 pa
• Permanent

Research, develop and evaluate image processing and computer vision algorithms within a specialist team, applying AI and deep learning to segmentation, object detection, tracking and classification. Work spans optical flow, pose estimation, particle filtering and computational imaging using MATLAB, Simulink and Python with PyTorch, collaborating with multidisciplinary engineers to turn research into practical applications.

MATLAB · Simulink · Python +1

• London
• Hybrid · 2–3 days in office
• Mid-level

Builds the simulation side of Large Physics Models, generating multi-fidelity CFD data to train large-scale ML models. Develops solvers and numerical methods in differentiable or GPU-native frameworks such as JAX or NVIDIA Warp, using commercial and open-source CFD tools for reference data. Works in a research team alongside ML researchers, shaping data generation strategy and fidelity tradeoffs.

JAX · NVIDIA Warp · OpenFOAM +2

• New York, United States
• Hybrid
• Lead

Lead the ML approach to manufacturing and MRO data, taking models from design through deployment in customer production environments across cloud, on-prem and edge. Work with sparse, drifting sensor and process data, fusing physics simulation with shop-floor measurement, and set uncertainty quantification and validation standards. Mentor engineers and data scientists while travelling to customer sites roughly 3-4 weeks per quarter.

Python · PyTorch · Pandas +3

• London
• Hybrid · 3 days in office
• Senior

Research engineers translate frontier research concepts into practical implementations, building and optimising state-of-the-art AI models such as Transformers, GNNs and diffusion models, plus the codebases, data pipelines and infrastructure for training and evaluation. They design and run experiments, analyse model performance, and advise on bringing models to production, working closely with research scientists in a cross-disciplinary team.

JAX · PyTorch · XLA +6

• London
• Hybrid · 1 day in office
• £260,000 – £630,000 pa
• Mid-level

Join the Pretraining team to develop the next generation of large language models, working at the intersection of research and engineering. You'll lead small research projects, design and analyse experiments on model architecture, algorithms, data processing and optimizers, and scale distributed training infrastructure across thousands of GPUs.

PyTorch · Kubernetes · GPUs

• London
• Hybrid
• Senior
• Permanent

Lead the technical architecture and roadmap for onboard localisation, designing and productionising state-estimation systems that fuse GNSS, IMU and wheel odometry for accurate vehicle pose estimates. The role spans algorithm design, prototyping, validation and deployment on real-time vehicle platforms, with hands-on technical leadership and mentorship across robotics, calibration, controls and sensor teams.

GTSAM · Ceres · g2o +7

• London
• Senior

Owns the core write-path operations for Delta Lake, the open-source storage layer behind the Databricks Lakehouse, driving performance innovations like Low Shuffle Merge and Deletion Vectors. Designs and maintains high-QPS, low-latency backend systems at massive scale, with technical leadership, mentoring and cross-team project planning.

Databricks · Delta Lake · Apache Iceberg +6