Staff Software Engineer, SRE

📋 Description

• • **Architect the backbone of developer velocity at Poshmark** – As a Staff Software Engineer on the Tools & Infra DevOps (SRE) team, you will own the design, implementation, and evolution of the CI/CD and non-production infrastructure that 300+ engineers rely on every day. Your work will directly accelerate feature delivery, reduce time-to-market, and raise the bar for operational excellence across our fashion-resale marketplace serving 130 million users.
• • **Lead transformational CI/CD initiatives** – You will re-imagine and scale our continuous-delivery stack (currently Jenkins/GitHub Actions/Spinnaker) to handle thousands of daily builds, tests, and deployments with sub-10-minute feedback loops. Expect to introduce advanced caching, parallelization, dynamic test selection, and self-service pipeline templates that any squad can adopt in minutes.
• • **Automate and harden non-production environments** – From ephemeral dev boxes to long-lived staging shards, you will codify every layer of infra with Terraform, Helm, and Ansible to guarantee parity, security, and rapid spin-up/teardown. You will also champion remote-development environments (GitHub Codespaces, dev-containers, or similar) so engineers code in reproducible, cloud-native sandboxes.
• • **Observability that predicts, not just reacts** – Design and roll out next-gen monitoring, logging, and tracing that cut MTTR by 50 %. Integrate SLIs/SLOs, synthetic checks, and anomaly detection into Grafana, Prometheus, and Honeycomb dashboards that surface actionable insights before users feel pain.
• • **DevEx evangelist and mentor** – Pair with feature teams to identify friction, then deliver turnkey tooling—build accelerators, automated canaries, golden-path templates—that let developers ship faster without becoming infra experts. Host guilds, brown-bags, and code-reviews to spread SRE culture and up-level peers.
• • **Cross-functional force multiplier** – Collaborate daily with Security to bake in secrets-management, policy-as-code, and vulnerability scanning; with QA to shift-left performance and integration tests; with Data Engineering to streamline ML model deployments; and with Platform teams to align on service-mesh, multi-cluster, and cost-optimization strategies.
• • **6-month wins you will deliver** – (1) Reduce average CI pipeline runtime by 30 % through build-graph optimization and runner autoscaling. (2) Achieve 99.9 % pipeline availability with automated rollbacks and health checks. (3) Cut environment-provisioning time from hours to <15 minutes via fully declarative IaC. (4) Establish a living observability runbook adopted by 80 % of squads.
• • **12-month vision you will shape** – (1) A self-service developer portal where any engineer can spin up a prod-like environment, run a full test matrix, and deploy to staging in one click. (2) A modern, cloud-agnostic CI/CD stack (GitHub Actions + ArgoCD or equivalent) with feature-flag-driven progressive delivery. (3) A culture where non-prod incidents are rare, measured, and resolved via blameless post-mortems. (4) Reusable frameworks (Terraform modules, Helm charts, policy packs) that become the gold standard across Poshmark.
• • **Strategic roadmap influence** – Use data-driven insights (DORA metrics, cost-per-build, dev-NPS) to prioritize quarterly OKRs, secure headcount, and justify investment in cutting-edge tooling. Present findings to VP-level leadership and shape the multi-year infrastructure strategy.
• • **On-call & incident command** – Serve as the senior escalation point for non-production incidents, leading war-room response, coordinating mitigations, and ensuring follow-up actions are owned to completion. Turn every fire-drill into a learning opportunity that hardens the platform.
• • **Tech radar & experimentation** – Run proof-of-concepts on emerging technologies—e.g., WebAssembly-based build artifacts, ephemeral Kubernetes clusters, eBPF for kernel-level observability—and publish internal tech-notes that guide adoption decisions.
• • **Sustainability & cost consciousness** – Balance performance with cloud spend; implement intelligent autoscaling, spot-instance orchestration, and resource-rightsizing that save six-figure dollars annually without sacrificing reliability.