Full-Stack Developer (AI-Native)

I'm the founder of two US software products and I'm hiring one full-time developer to work directly with me. No agency, no layers, no project managers.

The products:

Revphlo: B2B sales analytics for US sales teams (SaaS)
MoneyPilot: consumer personal finance app (fintech)

The job:
Requests come in through Slack all day: bug reports, small features, fixes, asks. You own that queue. You turn a vague Slack message into a scoped task, confirm it in one message, and ship it. As trust builds, you take bigger roadmap items: new endpoints, integrations, a security hardening pass, AI features.

The stack:
Next.js (App Router), TypeScript, React, Tailwind + shadcn/ui, Prisma ORM, Postgres on Neon, Clerk (auth), Vercel (hosting), Stripe, LLM APIs (Claude, OpenAI).

How we work:

AI-first shop. I pay for your Claude Code/Codex subscription and expect you to use AI aggressively for building, debugging, and testing. I also expect you to catch AI when it's wrong. If you don't trust AI tools or can't work without them, this role fits neither extreme.
No standup theater, no ticket bureaucracy. If you're stuck more than 2 hours, you post what you tried and ask for help, you will be helped immediately.

You:

3+ years building and shipping real production web applications (not just AI-assisted side projects)
Daily user of agentic AI tools (Claude Code, Cursor, Codex or similar)
Can debug a production issue from a vague report, across frontend, API, and database
Have built or experimented with features that call LLM APIs
Care about security basics: secrets handling, auth checks, input validation
Clear written English, proactive communicator
Available and responsive during US Eastern business hours (minimum 10am-4pm ET)

Pay:
$2,500-$3,000 USD per month depending on experience, full-time, paid reliably every month, plus a monthly reliability bonus. Long-term role, not a project.

To apply:
Start your application with the word SPEARGUN, then answer the 5 questions below. Applications missing the word or with generic copy-paste answers are deleted without reply.

1) Describe your AI development workflow in specific detail. Which tools, and walk me through how you would take a vague bug report ("the dashboard number looks wrong") all the way to a deployed fix.

2) Link to something real you built or heavily contributed to (repo or live product). What is the weakest part of it and why?

3) Have you shipped a feature that calls an LLM API in production? What did it do, and what's one problem you hit?

4) A founder sends you a one-paragraph spec that's missing details. What do you do, step by step? (Tests: plan-first behavior, communication.

5) Logistics: expected monthly salary in USD, hours you can commit to overlapping 10am-4pm US Eastern, earliest start date.