About the role

Role Overview

Unstructured is hiring a Site Reliability Engineer (SRE) for its small, technically deep Infra team. The team owns the reliability and performance of the platform end-to-end, working across infrastructure provisioning, Kubernetes operations, workflow orchestration, and core services.

What You’ll Own

Production reliability across Unstructured’s Knative/KEDA/Kubernetes document processing platform
Proactive detection and diagnosis of degradation, root-cause analysis, and shipping fixes
Observability & SLOs: end-to-end tracing, latency SLOs, capacity dashboards, and alerting
Load testing & capacity planning: define throughput benchmarks and detect performance regressions early
Fleet operations: contribute to safe, automated upgrade processes for production systems

What We’re Looking For

4+ years in SRE, platform engineering, or infrastructure engineering in a production Kubernetes environment
Strong Kubernetes operational expertise, including hands-on use of:
- HPA, KEDA ScaledObjects
- PodDisruptionBudgets, preStop hooks, PriorityClasses
- Understanding of pod lifecycle and scheduler behavior
Proven experience diagnosing and resolving production performance issues (e.g., resource saturation, timeouts, scheduling problems, graceful shutdown gaps)
Enough Python or Go to read service code, trace bugs to root cause, and implement targeted fixes

Nice-to-Haves / Implied Strengths

Experience with Knative and Kubernetes-based workflow orchestration
Comfort owning incident process maturity and reliability engineering practices

About Unstructured

Unstructured is an enterprise data transformation company focused on generative AI and LLM-ready data pipelines. Its open-source tooling is used widely across commercial and federal production workflows to convert documents and other content types (e.g., PDFs, HTML, Word, images, email) into scalable AI-ready outputs.