Senior Software Engineer, Compute

Persona

1d•$190,000 - $250,000

About The Position

Persona is the configurable identity platform built for businesses in a digital-first world. Verifying individuals and organizations is harder — but more important — than ever, with AI enabling fraudsters to launch sophisticated accounts at scale and regulations evolving rapidly. We’ve built Persona to support practically every use case and industry — that’s why we’re able to serve a wide range of leading companies. For example, Reddit relies on Persona for age assurance and verification to comply with online safety regulations, protecting younger users while maintaining a seamless experience. Meanwhile, OpenAI relies on Persona to keep bad actors out, protecting one of the world’s most powerful AI platforms from large-scale abuse in a time when AI is reshaping the way we work and live. We’re growing rapidly and looking for exceptional people to join us! The Compute team's mission: any engineer, using AI, should be able to stand up and operate a stateful service at production scale with no meaningful infrastructure knowledge required. We build the tiering zones, guardrails, and AI skills that make this possible. Reliable, observable, and self-service by design. This is a software engineering role. You'll spend real time writing code: platform APIs, internal tooling, deployment pipelines, automation that replaces the support tickets. Infrastructure fluency matters; infrastructure specialization is not the lead. What matters most is product instinct: understanding your users (product engineers), measuring whether they're getting unblocked, and treating developer experience as a feature acceptance criterion. The platform is yours to own, not just build. When ArgoCD fails and every team's deploy is blocked, that's your page. When a product engineer can't figure out why their service is OOMKilling and there's no useful error surface, that's your design failure to fix. A Senior engineer will help define the platform's interfaces and own how product teams interact with compute infrastructure. The team is still establishing its patterns, and you'll have real influence over what they become.

Requirements

Strong SWE fundamentals: shipped production services, writes readable and operable code, can debug distributed system failures across layers you don't fully own
Product instinct: you measure adoption, treat low usage as a design failure, and change things based on user feedback rather than your own assumptions
Developer experience sensibility: you identify friction in CLIs, APIs, and deployment workflows and design for clear affordances, actionable error states, and progressive disclosure
Kubernetes in practice: you've operated services on K8s, can debug scheduling failures, write RBAC policies, and work out why a pod won't start. You're willing to be paged for the platform you build.
Infrastructure as code in a team context: you've modified production IaC alongside other engineers and understand why declarative infrastructure and change safety matter
Senior candidates: we look for production-level Terraform ownership, including state, module composition, and clean infrastructure refactors.

Nice To Haves

Deep K8s internals: cluster lifecycle, admission controllers, custom controllers/operators
ArgoCD or another GitOps tool in production
GKE or managed Kubernetes at scale
Ruby or Rails (we integrate with Ruby systems)
Experience in a regulated or high-trust environment

Responsibilities

Build internal platform tooling and APIs that product engineers use to deploy, scale, and observe services, with developer experience as a first-class requirement
Own and evolve the Kubernetes platform, ArgoCD deployment pipeline, and Terraform IaC that underpin every service at Persona
Improve platform observability so "why is my service slow" and "why did my deploy fail" have self-service answers. Instrumentation is a platform primitive, not a per-service afterthought.
Drive Cloud Run-to-Kubernetes migration, working directly with product teams to remove blockers and hand off operational ownership cleanly
Design self-service workflows to reduce the Compute interrupt load. Success means product engineers can do more without filing a ticket.
Define the interfaces and abstractions the platform exposes to product teams. What does "easy to use correctly" look like for a deployment workflow or platform API?