SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
CuspAI is a frontier AI company focused on accelerating materials discovery through foundation models and AI. The company's mission is to unlock breakthrough materials needed for energy, clean water, advanced computing, and carbon capture—solving in months what nature took billions of years to perfect.
You will join a small, high-impact team as a Research Engineer (titled Member of Technical Staff) responsible for building and training specialized foundation models on terabytes of scientific data. This role spans multiple modalities: crystal structures, material properties, text, and images. Depending on your background, you may lead efforts in self-supervised representation learning over crystal structures, LLM fine-tuning and post-training for scientific tasks, or the evaluation frameworks that validate model performance.
Key responsibilities include:
- Designing and executing experiments at scale: scaling studies, ablations, and evaluations to determine which modeling approaches are most effective
- Leading parallel training of multi-modal foundation models across hundreds of GPUs, handling terabyte-scale datasets
- Applying strong software engineering practices to ensure systems are scalable, reliable, and maintainable
- Owning data curation and evaluation design end-to-end for your models
- Partnering closely with chemists and materials scientists to integrate computational and experimental workflows
- Learning materials science and experimental chemistry vocabulary to collaborate effectively with domain experts
- Ensuring all system designs align with CuspAI's sustainability commitments
You will own models end-to-end, from data and experiment design through training at scale to integration into the platform. The role is based in Cambridge with flexibility to work from London, Amsterdam, or Berlin offices, with an expectation of being in-office three days per week. Regular travel to other offices for collaboration may be required.
REQUIREMENTS:
- Master's or PhD in Computer Science, Machine Learning, or a related quantitative field, or equivalent professional experience
- Strong software engineering skills and proven track record building complex, high-performance systems in production or industrial environments
- Hands-on experience training or evaluating large models in distributed settings; for senior candidates, having led multi-hundred-GPU training efforts; for earlier-career candidates, core contribution to such efforts
- Deep expertise in at least one of: large-scale vision or representation learning, LLM pre-training or post-training, or reinforcement learning (depth in one substantial, scale-focused project preferred over breadth)
- Experience with representation learning and computer vision (self-supervised or contrastive pre-training, embeddings, models extracting structured information from images)
- Deep interest in the connection between AI and materials science or chemistry, with willingness to rapidly learn scientific vocabulary
- Personal motivation to enable scientists working on world-changing challenges in materials discovery
- Typical career stage: 1–3 years at a frontier lab for early-career candidates; senior and staff-level engineers with model program leadership experience also sought
BONUS QUALIFICATIONS:
- Experience with RL for post-training (RLHF, GRPO-style methods) or RL in agentic settings
- Background at a frontier lab or team shipping models at scale
- Prior ML applications to materials science, chemistry, or molecular informatics
- Exposure to multi-modal training and modality combination techniques
