Materials Science Jobs

• London
• Hybrid
• Permanent

Owns the full lifecycle of data-rich tools, services and web applications, building scalable ingestion, processing and cleaning pipelines for structured and unstructured materials-science data. Develops Python APIs and UIs for querying, exporting and visualising data, with orchestration, governance and containerised deployment using Docker and Kubernetes.

Dagster · Flyte · Prefect +6

• London
• Hybrid · 3 days in office
• Permanent

Owns the design of scientific data representations, schemas and ingestion pipelines that turn heterogeneous external materials data into ML-ready assets. Works with AI researchers and materials scientists to harmonise experimental and computational data from partners, defining quality, validation, lineage and provenance frameworks.

Pydantic · JSON Schema · pymatgen +7

• Cambridge
• Hybrid · 3 days in office
• Permanent

Trains and evaluates multi-modal foundation models for a materials discovery platform, spanning crystal structures, properties, text and images. Owns a model end to end, from data curation and experiment design through large-scale distributed training to integration, working closely with chemists and materials scientists.

GPUs · SFT · RLHF +1

• London
• Hybrid · 3 days in office
• Mid-level
• Permanent

Owns a set of partner labs and their integration onto an AI materials-discovery platform, translating experimental workflows, instruments and LIMS/ELN data into engineering requirements. Leads technical discovery and onboarding through to first closed-loop campaigns, and works with ML researchers on Bayesian optimisation campaigns that respect real lab constraints.

Python · LIMS · ELN +3

• London
• Hybrid
• Entry level
• Permanent

Develop fast, accurate machine learning force fields distilled into lightweight student potentials for high-throughput Monte Carlo simulation, then integrate them into the in-house kUPS framework. The role involves curating training datasets, running validation campaigns against classical force-field baselines, and profiling the MC inner loop for throughput. Embedded in the chemistry team, with a publication on MLFF-driven GCMC for MOF screening.

kUPS · MLFFs · Equivariant Models +2

• London
• Hybrid · 3 days in office
• Senior
• Permanent

Designs and builds agentic frameworks that orchestrate closed-loop scientific workflows, connecting agents to ML models, simulation engines and databases. Applies Bayesian optimization and active learning for experimental design, and builds evaluations for agent planning and decision-making. Works closely with chemists and materials scientists on customer projects.

PyTorch · JAX · LLM