Software Engineer, Infrastructure

About the Role

Platform owns the foundation Descript runs on: compute and deployment, reliability and on-call, CI/CD and monorepo health, developer environments, the infrastructure that model training and inference run on, and the security boundaries around all of it.

You will contribute to how models are trained and served, as well as how agents work inside our codebase: the environments they run in, the verification that makes their output trustworthy, and the review paths that keep it all legible.

What You'll Own

• Platform Ownership: GCP, Kubernetes, Temporal, the GPU fleet behind cloud export, and the deploy and rollback machinery everything ships through. Participate in the on-call rotation and make it quieter and more actionable.
• AI Enablement Substrate: GPU capacity, training and inference pipelines, and the reliability and cost of the systems serving models in production.
• Cost as an Engineering Constraint: Balance smart trade-offs; manage metering and attribution behind inference costs.
• Security: Manage identity and access, secrets management, least-privilege boundaries, and supply-chain integrity.
• Legibility & Observability: Infrastructure-as-code that explains itself, clear runbooks/repo context, and observability indicating failures and causes.
• Delivery & Standards: Form hypotheses, instrument work, release incrementally, and strengthen tooling, tests, and release practices.
• Team Growth: Provide architectural direction, thoughtful reviews, mentoring, and clear writing.

Qualifications

Required

• 8+ years building and operating production distributed systems, or equivalent server-side engineering with a heavy infrastructure focus.
• Effective use of agents and critical thinking about leveraging AI in your workflow.
• Proven experience operating systems where failure was expensive, with strong reliability and deployment practices.
• Experience carrying a pager, incident commanding, and managing rollbacks.
• Hands-on use of SLOs and error budgets as operating tools.
• Production experience with a major cloud provider and Kubernetes, using infrastructure-as-code by default.
• Experience owning an architecture or migration from planning through launch.
• Track record of identifying, scoping, and delivering unowned work without a rigid spec.
• Rigorous debugging approach (repro, log analysis, targeted verification checks).

Preferred Experience

• GPU and ML infrastructure: capacity planning, training or inference pipelines, serving cost and latency.
• Production security engineering: IAM, secrets, supply chain, least privilege.
• Cloud cost modeling: commitment strategy, reservations, unit economics.
• CI/CD at monorepo scale and developer-environment work.
• Experience with media, video, or GPU-backed workloads.
• Experience on small teams owning large surfaces (Series B to D startups or internal platform teams).

Application Note

Descript promises to review any application that includes a share link to a Descript video (about anything—career, favorite hobby, etc.) created using the product.