Member of Technical Staff, AI Supercomputing

About the Company

Radical Numerics is an AI research lab building general biological intelligence. Our mission is to master the code of life and reduce human suffering through generative genomics and large-scale biological world models.

Responsibilities

• Operate and automate large GPU clusters, including provisioning, imaging, and capacity planning.
• Build a unified compute interface to abstract cluster management for researchers.
• Extend scheduling and orchestration using systems like Kubernetes or Slurm.
• Maximize throughput and hardware efficiency by profiling and tuning performance across the stack.
• Improve reliability and recovery through monitoring, fault tolerance, and checkpointing.
• Design durable storage and artifact paths for datasets, checkpoints, and logs.
• Collaborate with model researchers to advise on parallelism and performance trade-offs.

Requirements

• Proven track record operating large-scale GPU clusters and container orchestration (Kubernetes or Slurm).
• Proficiency in Python or Rust for building performant software.
• Strong systems background in Linux, networking, and infrastructure-as-code.
• Understanding of deep learning frameworks and systems internals (PyTorch, Triton, CUDA, C++).
• Ability to debug complex distributed training and memory/performance regressions.
• Excellent communication skills to bridge technical and scientific domains.

Preferred Qualifications

• Familiarity with CUDA/NCCL and performance profiling.
• Experience supporting large-scale distributed training for frontier models.
• Contributions to open-source ML systems like PyTorch, Torchtitan, or Megatron-LM.
• Background in applied math, systems, computational biology, or quantitative sciences.