You will work with OpenMM, OpenFE, molecular dynamics, MM/GBSA, free-energy perturbation (RBFE/ABFE), and GPU infrastructure to define and validate automated biophysics tools. The team builds a scalable platform for agentic drug discovery workflows, ensuring scientific correctness in molecular simulation.

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.

Work with TensorFlow or PyTorch, Python, and related computer vision/graphics libraries on neural network architectures (convolutional, recurrent, transformer, diffusion-based). The Visual Computing Group (VCG) researches and develops generative models for game rendering and live video streaming to advance state-of-the-art runtime efficiency, visual quality, and latency.

You will work with backend development, databases, APIs, cloud services and production machine learning systems. This is for a growing technology business. No specific project or team size is mentioned.

The role involves developing embedded software and software for high-speed network platforms, working with GPU-based infrastructure and interconnect technologies. The company builds next-generation networking solutions integrating hardware and photonics for high-performance computing systems.

Hudson River Trading★4.5 (2)· London

You will work with Python, Pandas, NumPy, R, and MATLAB to apply statistical analysis and machine learning to vast quantities of market data. The team builds and maintains predictive trading models for the company's high-performance trading infrastructure. You will research, develop, and test order execution methods to increase trading efficiency.

Hudson River Trading★4.5 (2)· London

You will use Python, C++, Pandas, NumPy, R, or MATLAB to build and maintain predictive trading models. As an algorithm developer, you research and test novel order execution and model training methods, applying

You will use TensorFlow or PyTorch, Python, and CV/graphics libraries, plus 3D engines like Unreal or Unity, to develop generative models for graphics and video. As part of the Visual Computing Group, you'll design neural networks and ML for PlayStation's game rendering and live

Quantitative Researchers at Jane Street build models, strategies, and systems for pricing and trading financial instruments, working with Python, petabytes of data, and a large GPU cluster. They apply statistical and ML techniques, from linear models to deep learning.

DeepL★3.8 (4)· London

Work with Kubernetes, Terraform, ArgoCD, Go or Python, and AWS across hybrid on-prem and cloud infrastructure. The Hybrid Platform Engineering track builds and operates the compute layer—Kubernetes control plane, networking, storage, and core services—powering DeepL's language AI products and research. The role involves deepening the hybrid model and unifying workloads across on-prem and AWS.

You’ll train ML models, work on ML libraries and workflows, and use a GPU cluster with H100/H200/B200s. Projects are real-world machine learning applications for Jane Street’s global trading business, testing ideas against noisy financial data. No team size is stated.

Senior Machine Learning Engineer at ASOS, building production ML systems for personalized outfit recommendations and style discovery. You'll work with deep learning frameworks like PyTorch/TensorFlow, distributed GPU infrastructure, and MLOps practices including CI/CD, containerization, and observability. Collaborate with ML Scientists on batch and real-time models serving millions of customers.

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.

Work with Python, DevOps, Linux, Pandas, NumPy, and relational/non-relational databases. You'll join the Vendor Data Group at a global quant firm to build and maintain mission-critical ETL/ELT data integration pipelines and tools for alternative commodity and macro data, collaborating with a trading team.

Design, build, and optimize large-scale HPC/AI infrastructure using GPU optimization (CUDA, NCCL, GPUDirect RDMA, NVLink). You'll work with triple-digit petabyte-scale storage and massive CPU/GPU clusters, using Python, C/C++, and Linux in globally distributed data centers. This role covers HPC cluster design, performance tuning, and automation for thousands of nodes.

Develop end-to-end solutions in Python (pandas/numpy) for the Commodities RAD team, building data, research & trading platforms and tools for Portfolio Managers. Also use React/Angular for web frontends and SQL for databases.

Working with Core Python, Pandas, NumPy, Flask/FastAPI/Django, SQL, and Unix/Linux, you will build backend applications and data processing pipelines for trading and risk management within an investment banking environment, supporting risk calculations, trade workflows, and reporting.

Zanders★4.5 (2)· London

Work with modern C++ (STL, templates, RAII, unit testing, build systems); Python or Java; Linux; GPU compute (CUDA, ROCm); and HPC tools (MPI, Slurm). Lead quantitative engagements designing, developing, and validating models for market risk, counterparty credit risk (CCR), CVA/XVA, margin models, and derivatives pricing across asset classes. Act as a career coach for up to three direct reports.

Imperial College London★4.2 (238)· London

Work with PyTorch, GPU kernel programming, Verilog, and NVIDIA's Holoscan Sensor Bridge to integrate FPGA-based AI-ISP accelerators with edge GPU platforms. The project targets low-latency, high-throughput sensor-to-GPU streaming for the "Sense-Decide" pipeline in edge automation.

S&P Global★3.9 (1.7k)· London

Lead the AI+ strategy for S&P Global's Finance team, using AI/ML, NLP, and GenAI to improve finance processes and operational efficiency. This role will develop and execute the roadmap for automation and robotics, partnering with Corporate Finance and divisional CFOs. The team is within the Enterprise Finance and Strategy group.

Columbia Threadneedle Investments★3.7 (34)· London

This role develops Python-based automation solutions and builds Power BI dashboards, using tools like Microsoft Fabric, Copilot, and Power Automate. You will join the Controllership function within Finance, a global team of about 70, to drive automation, data analytics, and process improvement for record-to-report activities.

Work with C++, Python, ROS/ROS2, Linux/RTOS, and multi-core environments applying ISO 26262 and SOTIF. Execute hazard analyses, FMEA/FTA, and derive safety requirements for AeroVect’s autonomous ground-handling platform used by airlines. Collaborate cross-functionally to ensure safety integration and verification across supervised and unsupervised autonomy.

You will work with Python, PyTorch, JAX, and TensorFlow to develop optimisation, control, and machine learning methods including reinforcement learning and physics-informed ML. The role is in UCL Computer Science's Multi-Sensory Devices Group, whose research covers wearable sensing, robotics, and biomedical devices.

You will work with C++, Python, ROS/ROS2, Linux/RTOS, and static analysis tools to perform hazard and safety analyses. AeroVect is developing autonomous ground handling systems for airlines and ground service providers, covering perception, motion planning, localization, and vehicle control. This role executes safety lifecycle tasks under the Head of Safety for both supervised and unsupervised autonomy.

Lead engineers use Java, C#/.NET, Python, TypeScript, JavaScript, React, Go, Ruby, Scala, R, and SQL; Git, MacOS/Linux, and CI/CD; and AI-assistive tools. They lead multi-client software consultancy engagements, building and delivering production agentic systems and ML-powered features. The team is typically one to two clients per year.