Teams at CuspAI

Build and operate security controls for AI models, research data, cloud infrastructure, and CI/CD pipelines. Responsibilities include defending against data poisoning and prompt injection, implementing AI security posture management, securing Docker and Kubernetes environments, codifying cloud security policies with infrastructure as code, managing IAM and secrets, monitoring for incidents, and promoting security practices across interdisciplinary engineering and research teams.

Build and operate a cloud-native ML infrastructure platform on Kubernetes. Own MLOps tooling, CI/CD, model registries, experiment tracking, distributed GPU training, monitoring, configuration management, cost optimization, and deployment automation. Create developer tools that simplify the path from research experiments to production while collaborating with ML researchers, chemists, materials scientists, and software engineers.

Optimize large-scale AI training and inference workloads across NVIDIA GPU infrastructure. Profile model execution, identify bottlenecks, improve MFU, develop CUDA or Triton kernels, optimize distributed communication, streamline data pipelines, and build observability dashboards and scaling tools. Collaborate with AI researchers and infrastructure engineers to improve model-hardware compatibility and future cluster design. The role is based in European offices with three days per week onsite and regular international travel.