Descript
Software Engineer, Infrastructure
$220K–$292K
San Francisco, CA or Remote, US +2 · Remote · Full-time · Posted Sep 4, 2026
- Location
- San Francisco, CA or Remote, US
Remote
San Francisco, California, United States - Experience
- 8+ years, or equivalent server-side engineering with a heavy infrastructure focus.
About the job
Own and improve Descript’s infrastructure platform, including production reliability, deployments, GPU-backed model training and serving, security, and developer tooling.
What you'll do
- Operate and improve GCP, Kubernetes, Temporal, the GPU fleet, deployment and rollback systems, and on-call practices.
- Build and maintain infrastructure for model training and inference, including capacity, reliability, cost tracking, and production security.
- Guide platform architecture and improve CI/CD, observability, testing, documentation, and engineering practices.
What they ask for
- 8+ years building and operating production distributed systems, or equivalent server-side engineering with a heavy infrastructure focus.
- Production experience with a major cloud provider, Kubernetes, and infrastructure-as-code.
- Experience owning architecture or a migration from planning through launch.
- Experience with production reliability, including on-call, incident response, rollbacks, SLOs, and error budgets.
Nice to have
- GPU/ML infrastructure, production security engineering, cloud cost modeling, or CI/CD at monorepo scale.
- Media, video, or GPU-backed workloads; experience on small teams owning a large surface.
Pay
USD $220,000 to $292,000 base salary
Tools
GCP, Kubernetes, Temporal
AI focus
Contribute to AI agent tooling and the infrastructure for model training and production inference, including GPU capacity, pipelines, reliability, and cost.
Model-written summary · the original posting is authoritative