AI-Native Engineering Manager (Claude Code) — Launches, Bug Fixes & Delivery for a Shopify SaaS

About Shoperator Shoperator is a Shopify app for e-commerce merchants: AI ad, image, and video generation, a text-message "operator" that runs store tasks over SMS, Meta Ads integration, and a credits/billing system on Shopify and Stripe. Small founder-led team, paying customers, shipping to production most days using Claude Code and Codex. Why we're hiring We ship fast with AI coding agents. What we need now is someone who makes that shipping reliable and managed: own the release process, run bug triage, keep the backlog honest, and drive Claude Code sessions to close tickets end to end (spec → implement → verify on preprod → deploy → verify on prod). You are the person who makes sure nothing goes live half-verified and nothing that's been reported falls through the cracks. This is a hands-on role. You will run agent sessions and read diffs yourself. It is not a "coordinate other developers" PM job. What you'll own 1. Release management Weekly release cadence with a written pre-deploy checklist: in-flight background jobs, pending DB migrations, env flags actually set in the running environment, what else is riding along in the build. Post-deploy verification against the running container / shipped bundle (not the config file), with a short written proof per release. Release notes and a rollback plan for every prod deploy. 2. Bug intake and triage One queue (Linear). Every bug gets a severity, an owner, and a due date within 24 hours of being reported. Reproduce on the local Docker stack, write a self-contained spec, hand it to Claude Code or Codex, and verify the fix on preprod before it touches prod. Close the loop with whoever reported it (founder, support, merchant). 3. Driving Claude Code Write task specs that an agent can execute without you in the room: files involved, acceptance criteria, how to test, what NOT to touch. Run sessions using an orchestrator + subagent pattern, keep CLAUDE.md and agent definitions current, and enforce PR hygiene: tests green, lint on changed files, code review pass, no unrelated changes. Review agent output critically. We keep a written list of bug classes we've been burned by (async ORM access after commit, NOT NULL columns without Python defaults, feature flags that silently default on/off when an env var is missing, non-unique sort keys in paginated endpoints, etc.). You'll enforce that list in review and grow it. 4. Backlog and decision hygiene Consolidate work scattered across Linear, Trello, Slack threads, and session notes into one groomed backlog. Weekly grooming with the founder. Maintain an "open decisions" log with owner and due date so product decisions stop blocking engineering quietly. 5. Launch coordination Before any feature or pricing change flips on, confirm the landing page, App Store listing, billing rails, and support/persona copy all say the same thing. Run a launch checklist and a 24–48 hr soak on preprod with real usage. 6. Operational health First responder on uptime/paging alerts during your hours; maintain runbooks for the top 10 failure modes. Monthly hygiene sweep: stale branches, unpushed commits, diverged preprod/prod branches, disconnected store credentials, stale integrations. 7. Weekly written status One page: shipped, in flight, blocked, decisions needed, risks. Stack you'll work across Python / FastAPI, async SQLAlchemy, Alembic, Celery, Redis, Postgres · React / TypeScript on Vercel · Docker Compose local stack · GitHub Actions CI (~2.5k backend tests) · DigitalOcean prod · Shopify Admin API + Shopify Billing, Stripe, Meta Marketing API, SMS provider webhooks, image/video generation APIs (fal, Kling, Gemini) · Linear, Slack, Claude Code, Codex. Must-haves You use Claude Code daily to ship real code, not just chat. You can explain your setup: CLAUDE.md, subagents, plan mode, hooks, how you review what the agent produced. 3+ years shipping web software in a small team. You can read a Python or TypeScript diff and say "this is wrong, here's why." You have run production releases for a product with paying customers and can describe a rollback you've actually done. Comfortable with Docker, git beyond the basics (cherry-pick, rebase, handling diverged branches), CI logs, and server logs. Writes clearly and briefly. Async-first communicator. At least 4 hours/day overlap with US Central time. Nice-to-haves Shopify app ecosystem · Shopify or Stripe billing · Meta Marketing API / app review process · Celery + Redis operations · Codex or other coding agents · Linear administration · Twilio-style SMS products. Engagement details Part-time to start (15–25 hrs/wk), growing with results. Long-term. Hourly. Please quote your rate. Paid one-week trial: triage three open bugs, ship one fix through preprod to prod, and deliver a written pre-deploy checklist and post-deploy verification. We'll decide together at the end of the week. How to apply Start your proposal with the word PREPROD so I know you read this, then include: A short (5–10 line) story of a bug you fixed using Claude Code, including one thing the agent got wrong and how you caught it. Your pre-deploy checklist for a production release, as a bullet list. Your weekly availability and hourly rate. Generic proposals and proposals that don't include all three items will not be considered.