Member of Technical Staff, Infrastructure and Training Systems

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 by bringing the rigor of distributed systems and model architecture to the challenges of biology.

Responsibilities

• Design and scale distributed training systems for large-scale biological world models
• Maximize throughput and hardware efficiency through performance optimizations, custom kernels, and memory efficiency
• Build reusable training frameworks, internal libraries, and abstractions to improve reproducibility
• Improve reliability through fault tolerance, checkpointing, monitoring, and incident analysis
• Collaborate with model researchers and training scientists to identify bottlenecks and unblock experiments
• Adapt infrastructure to support new architectures, multimodal models, and long-context training

Requirements

• Strong engineering track record in distributed systems or high-performance ML infrastructure
• Proficiency in Python, PyTorch, Triton, CUDA, and C++
• Deep understanding of modern deep learning frameworks and their systems internals
• Ability to debug complex, multi-layered distributed training and performance issues
• Excellent communication skills to bridge technical and scientific domains

Preferred Qualifications

• Experience with large-scale distributed training for frontier or foundation models
• Contributions to open-source ML systems like PyTorch, Torchtitan, or Megatron-LM
• Familiarity with ML runtimes, compilers, numerics, and communication libraries
• Background in applied math, systems, computational biology, or related quantitative sciences