Senior AI Runtime & GPU Training Systems Engineer

Cacheflow; San Francisco (CA); On-site; USD 190,000 - 265,000; Full time

Benefits offered by this job

Equity

Annual performance bonus

Comprehensive benefits

Job summary

Databricks is seeking a Staff Software Engineer to contribute to the AI Runtime platform for large-scale GPU training. The role centers around driving the architecture and scalability of the managed training stack, ensuring fast and reliable operations.

Candidates should have over 10 years of experience with distributed systems and GPU training, along with a passion for mentoring and engineering excellence. This position is based in San Francisco, California.

Qualifications

• 10+ years of experience in large-scale distributed systems.
• Hands-on experience with distributed training frameworks.
• Strong understanding of training resilience patterns.
• Solid grasp of GPU performance fundamentals.
• Experience building multi-tenant platform products in the cloud.

Responsibilities

• Drive the architecture and evolution of AIR.
• Solve complex problems in large-scale training.
• Push GPU efficiency and training performance.
• Build resilience for multi-node jobs.
• Collaborate with product, research, and platform teams.

Skills

Distributed systems

GPU training infrastructure

High-performance computing

ML systems

Distributed training frameworks

Communication skills

System design

Mentoring engineers

Education

BS in Computer Science or related field

Tools

PyTorch

FSDP

DeepSpeed

Megatron

Job description

Databricks is seeking a Staff Software Engineer to contribute to the AI Runtime platform for large-scale GPU training. The role centers around driving the architecture and scalability of the managed training stack, ensuring fast and reliable operations.

Candidates should have over 10 years of experience with distributed systems and GPU training, along with a passion for mentoring and engineering excellence. This position is based in San Francisco, California.