Description
Radical Numerics is hiring a Member of Technical Staff, Infrastructure & Training Systems to design and build distributed training infrastructure for large-scale biological world models. The role focuses on scaling training across distributed compute, optimizing throughput and hardware efficiency, developing reusable training frameworks, improving fault tolerance and observability, and collaborating with researchers and engineers. It requires strong distributed-systems and high-performance machine-learning infrastructure experience, proficiency in Python, PyTorch, Triton, CUDA, and C++, and the ability to debug complex training systems.
