You will use PyTorch, JAX, and scikit-learn to build ML models in low-data regimes for drug discovery and molecular optimization. You will develop active learning, meta-learning, and uncertainty-aware approaches, designing data acquisition strategies that guide experiments, simulations, and compound selection.

This role involves working with Python, PyTorch, JAX, CUDA, Linux, Docker, Kubernetes, Slurm, Ray, Flyte, and Argo. The team builds ML and physics-based infrastructure for drug discovery, including simulation workflows and agent-usable scientific pipelines. Engineers improve code quality, GPU efficiency, and distributed execution while making research tools scalable and maintainable.

Lead a team of Data Scientists and ML Engineers building AI solutions for financial markets using LLMs, Generative AI, Deep Learning, Transformers, Agentic AI, Python, PyTorch, TensorFlow, Hugging Face, LangChain, Semantic Kernel, Azure AI Foundry, AWS AI Services, and MLOps/LLMOps practices.

Design, build, and operate a Kubernetes-based MLOps platform using Helm, Terraform, Kubeflow, KServe, vLLM, and JupyterHub. The platform supports the full model lifecycle—experimentation, deployment, and scalable inference—for AI, ML, and LLM workloads in production.

You'll build and operate a Kubernetes-based AI/ML platform using Helm, Terraform, Kubeflow, KServe, vLLM, and JupyterHub. This role focuses on creating infrastructure, tooling, and guardrails so MLOps and data science teams can reliably deploy and serve models at scale.

Use deep learning frameworks, distributed computing, GPU-based training, and MLOps to build and deploy batch and real-time ML models. The Search & Recommendations team develops personalised recommendations, product ranking, and AI styling experiences for ASOS's online fashion platform, helping customers discover products.

Lead 6-7 engineers on core ML infrastructure for a systematic quant fund: distributed model training, LLM hosting/fine-tuning, and scalable deployment. Use Python with working knowledge of C++/Java. Own technical direction for real-world, high-stakes systems.

Waymo★3.0 (61)· London

You will use Python, deep learning frameworks (JAX, TensorFlow/PyTorch), and C++ (preferred) to develop data-selection algorithms and scaling laws for foundation model training. The AI Foundations team builds machine learning solutions for autonomous driving, focusing on reinforcement learning, generative modeling, and robust evaluation to improve the Waymo Driver. You will report to a Senior Staff Software Engineer.

This is a partnerships and business-development role, not a hands-on engineering position. You will work directly with founders, ML researchers, and scientists to create a global data-acquisition strategy for the company’s AI biology models. You will identify, negotiate, and secure complex data partnerships with health systems, biobanks, academic groups, and pharmaceutical companies.

Google★4.3 (6.2k)· London

Manage engineers building Google’s Learning Platforms within the Global Business Organization, which powers Google’s ecosystem of products and services for customers and partners. Work with back-end languages (Java, Python, Golang, C++) and front-end (JavaScript, TypeScript, HTML, CSS), plus ML model fine-tuning and evaluation.

Work with Python, PyTorch (or JAX/TensorFlow), RL fine-tuning frameworks like TRL or verl, and distributed multi-GPU training. The Reinforcement Learning Team advances RL, Bayesian optimisation, AI agents, LLMs, and vision-language models, applying them to AI for science, chemistry, physics, and robotics while publishing at top venues.

You'll work with Python, TensorFlow, PyTorch, TFX, Kubeflow Pipelines, and model serving technologies like TensorFlow Serving and Triton on machine learning systems for personalization, recommendation engines, and ranking algorithms. At Fractal, a strategic AI partner to Fortune 500 companies, you'll design and train ML models, build data pipelines for large-scale datasets, deploy models to production, and lead A/B testing and experimentation across a multidisciplinary team.

This role develops generative AI applications using Java, Python, React or Angular, LangChain, LlamaIndex, Hugging Face, Docker, Kubernetes, and AWS, with tools like GitHub Copilot. You will build scalable, AI-powered solutions for Citi’s global finance operations.

You will own a gradient-boosted ranking model (LambdaMART) benchmarked against a two-tower retrieval comparator, working in Python on AWS. The team is building a creator-to-product matching model and a multi-modal content model (video, speech, audio) trained on first-party affiliate conversion data to rank creators by actual sales, not follower counts. You will lead this R&D project for twelve months as the second of three technical hires.

Work with Python, PyTorch, diffusion models and flow matching. Join an interdisciplinary team deploying, adapting and optimising generative models for real-world scientific and industrial applications at an AI research company developing models for scientific discovery.

Work with Nvidia GB200 72NVL CX8 and AMD MIxxx architectures, coding in C, C++, C#, Java, JavaScript, or Python to benchmark, profile, and debug GPU clusters. You will help build foundational models powering Microsoft’s personalized AI assistant, Copilot, spanning training and inference infrastructure, model architecture, and data curation.

Work with Python, PyTorch, ONNX, TensorRT, Docker, cloud environments, MLFlow, and CI pipelines to build scalable training and inference systems. The role develops, optimizes, and deploys state-of-the-art 3D computer vision models for volumetric CT data on Qureight’s AI-powered imaging platform.

You will work with Python and JAX, contributing to the Gemini large-scale pretraining project at DeepMind. Responsibilities include improving training workflows, tooling, and codebase to support research and address issues in large-scale runs.

Research, develop, and evaluate state-of-the-art deep learning models for volumetric CT data using Python and PyTorch. You will contribute to self-supervised pre-training of large 3D foundation models and specialize models for segmentation, classification, and regression to support clinical-trial and real-world evidence programmes. The role is within the Machine Learning function at Qureight, which builds an AI-powered imaging platform for lung and heart disease.

DeepMind★5.0 (2)· London

You will work with Python and ML libraries such as JAX to solve time-sensitive issues in large-scale training runs. You will improve technical workflows and tooling across Gemini large-scale pretraining, a core DeepMind project advancing AI development.

White & Case★4.0 (120)· London

Works with Python, JavaScript, MCP, agentic AI frameworks, LLMs, and cloud platforms (Azure, AWS) to architect and deploy AI tools. As technical lead of the Applied AI Lab, the role oversees design, development, testing, and deployment of AI solutions to support legal practice and firm-wide innovation. Manages and develops a multidisciplinary team of legal engineers.

You will work with Python (NumPy, SciPy), C++, large datasets, distributed systems, and high-performance computing concepts. The team builds high-performance trading and analytics platforms for a HFT leader that transforms sports data into competitive insight, and you will convert prototype models into production systems and optimize for low latency and scalability.

You will work with Python, JavaScript, Model Context Protocol (MCP), agentic AI frameworks, LLMs, and cloud platforms (Azure, AWS). As technical

Work with Python, JAX, and ML/scientific libraries on Gemini large-scale pretraining. The team enables new research, responds to time-sensitive issues in large-scale training runs, and improves tooling and workflows for model architecture and data pipelines.

Build high-performance trading and analytics platforms using Python (NumPy, SciPy) and optionally C++, with large datasets, numerical computing, distributed systems, and high-performance computing. The team turns complex sports data into actionable insight for a HFT leader.