Databricks
Staff Software Engineer, AI Runtime
$190K–$265K
Mountain View, California +2 · Staff · Full-time · Posted Jun 8, 2026
- Location
- Mountain View, California
San Francisco, California
Mountain View, California - Experience
- 10+ years in large-scale distributed systems; significant depth in GPU training, high-performance computing, or ML systems.
- Education
- BS in Computer Science or a related field; MS or PhD preferred.
About the job
Leads architecture and engineering for Databricks AI Runtime, a serverless platform for large-scale GPU model training, improving its reliability, performance, and experience for developers and agents.
What you'll do
- Shape AI Runtime architecture and long-term technical direction, including reliability, observability, and readiness for new GPUs and workloads.
- Solve multi-node training challenges and improve training efficiency across models and accelerators.
- Lead cross-team projects with research, product, and customers; mentor engineers and champion engineering excellence.
What they ask for
- 10+ years in large-scale distributed systems, with significant depth in GPU training, high-performance computing, or ML systems.
- Hands-on experience training large models with FSDP, DeepSpeed, or Megatron, and fluency in parallelism strategies.
- Expertise in GPU performance, including accelerator architecture, high-speed interconnects, and kernels.
- Strong understanding of training resilience patterns, such as checkpointing, failure detection, and automatic recovery.
Nice to have
- MS or PhD.
Pay
$190,000 — $265,000 USD (base salary)
Tools
FSDP, DeepSpeed, Megatron
AI focus
Builds and improves infrastructure for large-scale GPU model training, including training reliability, performance, orchestration, scheduling, data loading, and checkpointing; AI agents also launch experiments.
Model-written summary · the original posting is authoritative