QA & Support Engineer

01 Mission

• Triage incoming tickets in Pylon across chat, email and shared Slack channels — resolving what doesn't need an engineer and communicating clearly with paying customers, including strategic accounts.
• Investigate generation errors with our observability stack (Better

  Stack, Cloud

  Watch, Sentry, Post

  Hog), reproduce when needed, and spot patterns across tickets to raise quality issues and feed the support knowledge base.
• Test and validate new features on staging before they ship — web app, cloud API, ML layer and SDKs — covering happy path, billing, rate limits, model variations and edge cases; your eye on staging is what unlocks releases.
• File precise, reproducible bug reports in GitHub against the right component: with logs, request IDs and clear steps to reproduce.
• Write and maintain automated test suites: end‑to‑end (Playwright), API smoke tests and regression scripts.
• Contribute small features and fixes: submit PRs, write or update tests.
• Improve monitoring coverage: add alarms, dashboards and synthetic checks for the areas you find brittle.

02 Scope of ownership

Owns

• Customer support: Pylon triage, ticket resolution, customer communication and pattern detection.
• QA on new features: staging validation across the stack, PR review and bug reports.
• Light development and test automation: end‑to‑end suites, API smoke tests, small features and fixes, and monitoring coverage.
• Ensure cohesion across the team and roadmap: cross‑team coordination and PR reviews.

Does not own

• Software Architecture decisions (Engineering)
• Infrastructure and deployment pipeline (Engineering)
• Product roadmap and prioritization (Product)

03 What we look for

• Two to four years in QA, support engineering or as a junior developer, ideally at a SaaS startup.
• Comfortable reading TypeScript or Python well enough to navigate a codebase, file a useful bug and ship a small patch with AI assistance.
• Strong written English and clarity under pressure when a customer is unhappy, with a bias for action — you don't wait for permission to investigate, reproduce or improve.
• Practical with common support and observability tooling: Pylon (or Zendesk / Intercom), Notion, Slack, Git

  Hub, Sentry, Better

  Stack / Datadog, Post

  Hog, Playwright / Cypress and the AWS console.
• Familiarity with AI coding assistants (Claude Code, Cursor, Copilot): we use them daily.
• Experience with image / video models (Stable Diffusion, Flux) or LLM APIs.
• Exposure to incident response, on‑call rotations or SOC 2 / compliance contexts.

04 Disqualifiers

• Cannot read code: not comfortable navigating a TypeScript or Python codebase even with AI assistance.
• A passive support agent who waits for permission to investigate, reproduce or improve.
• No interest in automation; treats testing as purely manual and repetitive.
• Expects a pure QA role with no customer contact, or a pure support role with no engineering contribution.

05 How we hire

• Intro call (30 min)
• Product session (60 min)
• Build live (60-90 min)
• Written POV (Async, 48h)
• References (Parallel)
• Founder conversation (45 min)