CuspAI develops artificial intelligence systems that help scientists discover new materials. Its platform brings together AI models, scientific data, simulation tools and laboratory work, with applications in semiconductors, energy, climate technology and advanced manufacturing.
- Build AI agents for autonomous materials discovery
- Work across applied machine learning, agent systems and reinforcement learning
- Join a Cambridge AI startup that raised $450 million in July 2026
The Cambridge-based company raised $450 million in Series B funding in July 2026. The round valued CuspAI at $2.6 billion and included backing from the UK government and Jeff Bezos’ investment fund.
The role
CuspAI is hiring an Applied AI/ML Engineer to design and build intelligent agents for its autonomous materials discovery engine. The role focuses on turning modern agent and reinforcement learning ideas into systems that can support real scientific work.
You will work with researchers and engineers to create agents that can reason across scientific tasks, use tools, plan experiments and improve through evaluation. The job is practical and product-focused, so strong ideas need to become dependable software rather than stay as research prototypes.
Likely areas of work include:
- Designing agent workflows for complex materials science problems
- Building and testing reinforcement learning or agent-based systems
- Connecting models with scientific data, simulation tools and internal services
- Creating evaluation methods that measure useful scientific progress
- Improving reliability, speed and scalability as experiments move into production
- Working closely with machine learning researchers, software engineers and scientists
Who CuspAI is looking for
This opening is best suited to an experienced applied machine learning engineer who has built agent systems, reinforcement learning applications or other complex AI products. You should be comfortable writing production-quality code and working through uncertain problems where the right technical approach is not obvious at the start.
Experience with scientific machine learning, foundation models, tool-using agents, distributed systems or materials science would be valuable.
Why apply
CuspAI is working on a less crowded side of the AI market: using models to find materials that could improve chips, batteries, clean energy and manufacturing. This role gives an applied engineer the chance to build agent systems around measurable scientific outcomes.
The position can be based in Cambridge, London, Amsterdam or Berlin. CuspAI expects employees to work from an office three days per week, with occasional travel between locations.