> Markdown version of [/jobs/ext/3970122-staff-software-engineer-ai-runtime](https://www.wearedevelopers.com/jobs/ext/3970122-staff-software-engineer-ai-runtime). Every page supports `.md` or `Accept: text/markdown`. Links point to the HTML versions so they work for humans too. Agent guide: [/agents.md](https://www.wearedevelopers.com/agents.md).

---

# Staff Software Engineer, AI Runtime

- **Company:** Databricks
- **Location:** San Francisco, CA, United States
- **Experience:** Expert
- **Salary:** $228,600.0 - $342,800.0
- **Contract:** Permanent contract
- **Skills:** Artificial Intelligence, Extract Transform Load (ETL), Distributed Systems, High Performance Computing, DeepSpeed, Information Technology, Machine Learning Operations, Megatron, Databricks
- **Published:** October 27, 2026
- **Apply:** https://startup.jobs/senior-staff-software-engineer-ai-runtime-databricks-10561676

## About the Role

* 15+ 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.
 * A background in building and running multi-tenant cloud platforms as products.
 * A track record of setting technical direction across teams and seeing it through to business impact.
 * The credibility to bring teams to agreement on difficult trade-offs, and a reputation as a mentor senior engineers seek out.
 * BS in Computer Science or a related field (MS or PhD preferred).

Pay Range Transparency

## Description

* Drive AI Runtime's architecture and lead Databricks' long-term technical direction in AI training infrastructure.
 * Define the reliability and observability strategy for a fleet that keeps growing in scale and complexity.
 * Shape AI Runtime's developer experience for people and agents, so even the largest training jobs feel as simple as a first experiment.
 * Anticipate industry trends and choose the technical bets that prepare AI Runtime for new GPU generations and workload types.
 * Take on the hardest multi-node training problems yourself, from orchestration and GPU scheduling to data loading and checkpointing.
 * Make training faster and cheaper by raising efficiency across model architectures and GPU accelerators.
 * Work with research, product, and customers to decide what we build next and align teams on hard technical calls.
 * Champion engineering excellence, raise the bar, and help us hire and grow great engineers.