Machine Learning Jobs in Woking

Synthesia · London

Hands-on principal IC role owning the ML platform that lets researchers and product teams train, serve and deploy generative models. Work spans research infrastructure, production serving, internal tooling and increasingly agent-oriented automation, with a focus on reliability, scalability and GPU/cloud cost efficiency. Not a pure ML engineer role; strong generalist systems skills in Kubernetes, Linux and Python are central.

ElevenLabs · United Kingdom

Builds and operates core platform foundations covering workspaces, user management, API keys, login, analytics and admin tooling, or generic resource and permissioning services used by product teams. Owns features end to end from backend services to settings UI, with safe rollout and migration of live, business-critical systems serving millions of users.

Ocado · Erith

Lead a team of engineers, technicians and trainees at an automated Customer Fulfilment Centre, allocating workload, developing staff and handling disciplinary matters. Requires hands-on mechanical and electrical engineering competence to make decisions across a 24/7 operation, working 12-hour day, night and weekend shifts.

Ocado · Monroe, Ohio, United States

Provide first response to engineering fault conditions and carry out electrical and mechanical planned and reactive maintenance across a highly automated customer fulfilment centre. Diagnose breakdowns, perform assembly and sub-assembly repairs from blueprints and schematics, analyse MHE performance data and complete quality inspections. Initial 6-9 months involve training and travel to another North American site, with shift work covering days and weekends.

ElevenLabs · United Kingdom

Join the research team at an AI audio company, training state-of-the-art models across modalities and improving them through post-training, data curation and architectural changes. You'll design rigorous benchmarks and evaluations to judge whether new model iterations are genuinely better, working autonomously in lean, high-velocity teams.

Amazon · London

Hands-on analysis and modelling of large multimodal datasets to measure and improve Rufus, Amazon's conversational shopping assistant. Designs and analyses A/B tests, builds predictive models and metrics dashboards, and works with NLP, information retrieval and recommender systems alongside scientists and engineers in a London team of around 150.

Relation Therapeutics · London

Builds and scales generative and predictive models of cellular behaviour, acting as the bridge between ML theory and implementation. Day to day involves optimising large neural networks, distributed multi-GPU/multi-node training, profiling GPU workloads, and productionising research code for multi-modal biological data.

Relation Therapeutics · London

Lead the platform engineering team owning hybrid on-prem and cloud compute, Kubernetes clusters, and MLOps infrastructure such as Kubeflow notebooks, GPU training and model serving. Set architectural direction, mature CI/CD, observability and security controls, and ensure scientific compute meets the needs of ML and data science teams. Reports to the VP of Engineering in a matrixed TechBio environment.

OpenAI · London

Designs, builds and maintains production harm-detection systems using SQL, Python, machine learning and language models, developing signals, classifiers and pipelines on shared engineering platforms. Owns protection strategy and cross-functional delivery for a harm area or product launch, connecting detection with policy, review, enforcement and measurement. Includes on-call rotation and exposure to sensitive content.

OpenAI · London

Builds and operates shared multi-cloud compute platforms behind ChatGPT and the OpenAI API, spanning Kubernetes cluster lifecycle, upgrades, patching and compliance. Develops reusable platform tooling, infrastructure automation, capacity management and autoscaling, with end-to-end production ownership and on-call incident response.

OpenAI · London

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 unsafe tool use. Highly cross-functional, partnering with core training, product, policy, legal and trust and safety teams to ship safety improvements into production models.

Quantexa · London

Builds and enhances scoring and detection tooling on Quantexa's Decision Intelligence platform, turning complex interconnected data into insights for financial crime detection. Works across the stack with agentic AI tooling, taking features from concept to release alongside product and delivery teams, with a growing stakeholder management focus.

Amazon · London

PhD students join an Amazon Science team to own an end-to-end research project, from scoping with a mentor to deploying algorithms, models or systems into production. Projects span computer vision for Prime Video quality, multimodal AI for Rufus, automated reasoning and formal verification, observability for distributed systems, and recommendation models. Interns may publish at top-tier conferences or see their work ship before the internship ends.

Amazon · London

Hands-on data science role focused on Alexa for Shopping, building and measuring multimodal conversational systems powered by large language models, information retrieval, recommender systems and knowledge graphs. Day to day involves analysing huge multimodal datasets, designing and analysing A/B tests, building predictive models and metrics dashboards, and working with applied scientists and engineers to move insights into production.

Amazon · London

Lead end-to-end delivery of complex AI/ML engagements for enterprise customers, from strategic planning through pre-production deployment and optimisation. Architect solutions on AWS services with a focus on Generative AI using Amazon Bedrock and SageMaker, while mentoring junior consultants and supporting pre-sales scoping. Based in AWS offices in London, Manchester, Bristol or Cambridge when not at customer sites.