Support & Deployment Engineer (JS/Typescript, PostgreSQL/AI coding tools) - FL Fulltime - NA/LATAM

The lead engineer writes the architecture and the build prompts. Someone still has to run the prompts, fix what breaks, test it against real data, keep small end-to-end builds moving through staging, and make sure the people using these tools actually know how to use them. Right now that queue has no dedicated owner. This role is that owner.

What you’ll do

• Fix bugs. Own the day-to-day bug queue across Compass, Atlas/Morgan, and related tools - triage, reproduce, fix, verify, close.
• Follow the build flow. Leads write the .md build prompts and spec docs. You execute them faithfully - paste into Claude Code or your tool of choice, review what comes out, and finish what the AI missed.
• Test before it ships. Every small build gets exercised against real scenarios in staging before it goes further. If it’s not tested, it’s not done.
• Deploy small E2E builds. Take a scoped feature from a written .md flow through to a working, deployed staging build - end to end, without hand-holding on every step.
• Train and support staff. When Compass or Atlas features go live, you’re the one who helps staff actually use them - quick trainings, walkthroughs, answering the “how do I” questions.
• Write it down. If you fix something twice, it becomes a note in the .md flow so the next person (or the next AI run) doesn’t hit it again.
• Flag what’s actually broken vs. what’s just unclear. Leads need to know the difference - a bug report and a confused user are different problems, and you triage which one you’re looking at.

• 1-3 years shipping production software, or equivalent hands-on project work you can walk us through.
• Comfortable reading someone else’s spec (a written build prompt or .md flow doc) and executing it without needing it rewritten for you.
• Working knowledge of TypeScript/JavaScript and Postgres, or the clear ability to pick both up fast.
• Daily user of an AI coding tool - Claude Code, Cursor, or equivalent. Show us something you shipped with it.
• Comfortable with Git, PRs, and staging-first deployment discipline. Nothing goes to production without sign-off - you respect that line.
• Can explain a technical thing to a non-technical staff member without talking down to them. Training is part of the job, not an afterthought.
• You say so when something’s broken, including when you broke it. No blame-shifting, no silent workarounds., * Experience with Supabase (or Postgres + a similar BaaS) and Railway or a comparable deploy platform.
• Any exposure to QA processes - writing test cases, regression checklists, or Playwright/Cypress.
• Prior support, helpdesk, or training role - you’ve explained software to people who didn’t build it before.
• Background in education, training, skilled trades, or veteran-serving organizations.
• Military, first-responder, or service connection.

• 30-minute screen with Jimmy - walk through a recent bug you fixed and a build prompt you executed, start to finish.
• Short paid take-home - we hand you a real .md build flow, you execute it and push to a staging branch. We pay for your time.
• Technical walkthrough of the take-home - what you changed, what you’d flag back to the lead, and why.
• 30-minute culture conversation with Tim.
• Offer extended within 7 days of first screen.