AI Systems Engineer - Forward-Deployed Builder

(This role plugs into our highest-need seat — product/app, sales-ops, or CS-ops systems.)

In one line: We hand you a messy, ambiguous, high-stakes problem — and you come back with a working, tested system. You think about the business and the user, not just the code.

Why this is one of the best seats in AI right now
We're a fast-moving AI company building multiple products at once — an AI app builder, ad-creation tooling, cold-outreach/GTM systems, and an internal operations platform that runs the whole business. We move fast, ship constantly, and we give builders the keys.

What you get here that you won't get most places:

• No budget cap for tokens and no babysitting. Best-in-class tooling, frontier models, every AI coding harness and agent framework you want — and nobody counting your tokens. If you've ever been throttled by a $20 plan while doing real work, that ends here.
• A real engineering org and a brilliant CTO to build alongside. You'll be in the trenches with people who are genuinely excellent — and you'll get better fast. People who join us tell us they learn more in months than they did in years elsewhere.
• Frontier work, not maintenance. Agentic systems, internal platforms, GTM automation, net-new products. You'll build with the newest models and harnesses the week they drop.
• Autonomy and ownership. Drop-a-brief-get-it-built. You own a domain end-to-end; you don't wait for a ticket.
• Top-of-market pay, paid properly and paid on time. We pay our best builders what they're worth, and we don't play games with it. (See comp below.)
• No politics. Speed, output, and quality matter more than process theater. We're direct with each other — genuine, but not soft.

What you'll actually do
You'll be embedded where the hardest systems problems are. Depending on the seat, that looks like:

• Deconstruct messy real-world workflows — sales ops, marketing ops, product ops, fulfillment, support — into their underlying steps, then build the system that makes them run automatically and correctly.
• Build agentic systems that ship value — agents, automations, internal tools, data pipelines, reporting/attribution — that real people and real revenue depend on.
• Build product — stand up and rebuild complex products fast (think: an AI app builder, a reporting dashboard, an internal ops platform), making sharp build-vs-buy-vs-reuse calls under a deadline.
• Own the unglamorous plumbing — the integrations, the data model, the tracking, the thing that has to actually work — that everything else is gated on.
• Reason commercially — figure out why a funnel, campaign, or product isn't working from first principles, and fix the real cause, not the symptom.

This is not static work. The systems need someone who keeps improving them — not someone who builds once and walks away.

Who you are

• You're obsessive in the healthy way. When you get into something you go all the way down — best resources, deepest forums, you can't help it. You have side-builds you're proud of, built because you wanted to, not for a job.
• You ship absurdly fast. You build instead of debating. You've turned around things in a day or a weekend that other people quote in months.
• You're an agentic-engineering native. You don't just use AI — you build with it at the systems level: agents, memory, tools, orchestration, evals, failure modes.
• You see the whole board. You're technical, but you reason about the business and the user too. You can take a fulfillment gap and see the revenue model, not just the bug.
• You're right the first time. "Probably fine" isn't in your vocabulary. When you hand something off, it's tested and it works. Someone shouldn't have to check whether your work is 50/50 wrong.
• You're high-agency. Drop you a one-line brief and you make sensible decisions, document your assumptions, and ship — you don't wait to be managed.
• You can take a punch and throw one. We're blunt. We'll attack your approach hard — because we attack each other's approaches hard, that's how the work gets good. You engage, you defend your thinking with substance, and you bounce back. You can hang in a room with direct Americans, dry British roasting, and EU bluntness without getting precious.
• You live in the frontier. Your feed is AI all day. You can name what you started using in the last two weeks and why.
• You're intellectually honest. You say "I don't know" instead of bluffing, and you trace problems to root cause instead of grabbing the first plausible answer.

Signals we screen for (you don't need all of them — show us the strongest)

• Evidence you build things you drive — agents, automations, tools, a product you started, or something significant you initiated and pushed through at a job. This matters more than your résumé or your degree. (Some of the best people here have non-CS backgrounds.)
• Hands-on depth with agentic / AI engineering — building with LLMs and agent frameworks, not just prompting a chatbot.
• Experience in the info-product, agency, or DTC/GTM world — you understand funnels, offers, paid ads, cold outreach, CRO, or how an agency actually makes money.
• GTM engineering / automation — n8n (strongly preferred), or Make/Zapier with genuinely complex builds (branching, error handling, API calls — not 3-step zaps).
• Comfort with APIs, webhooks, JSON, and a real coding language (JS/Python/TS). You can read and debug a system, not just generate it.
• Bonus: martech / sales-tech depth — CRMs (Close a plus), tracking/attribution, ESP/email systems, Whop/Stripe, multi-tenant platforms.

What this role is NOT

• Not a prompt-engineer-who-tinkers. We need someone who ships systems real revenue depends on — not someone who plays with AI toys and never lands value.
• Not a pure spec-taker. If you can only build exactly what you're told and can't reframe the problem, this isn't it.
• Not a manager/coordinator role. This is a hands-on building role — the person who makes the machine actually work.
• Not a place that needs all-nighters. We want you thinking at your best. Healthy obsession, not grind-cope.

Logistics

• Comp: For the genuine article this lands around $12k–$22k/month (paid monthly, paid on time, no games), with real upside on what you own. We don't lowball the real ones and we move fast on the number. International? A strong USD monthly rate goes a long way — we pay for the talent tier, not the postal code.
• Type & framing: We start as a paid engagement because the best builders we hire run their own things, and we respect that — it grows into whatever's right for both of us (bigger scope, a real stake in what you own). This is a peer seat, not a probation.
• Location: Anywhere with enough timezone overlap to actually collaborate with the team day-to-day. Remote.
• Exclusivity: Flexible to start; the strongest engagements grow into something primary over time.

How to apply
There's no long form and no unpaid take-home busywork. We want to see what you've built and how you think — a tight Loom (keep the recording under ~15 min; think as long as you want):

1. Show us something real you built — ideally a link to something live we can poke at. Your best work where you owned the hardest part and drove it: a side-project, your own product, an agent/automation — or something you initiated and pushed through at a job (we respect confidentiality; redact or whiteboard it). Walk us through the hardest problem and where it still breaks.
2. A rebuild teardown:"If you had to rebuild an AI app builder — pick one: rork.com, Lovable, or Base44 — from scratch in 7 days, how would you attack it?" Hit the architecture, where you'd use agents vs. code, what you'd build vs. reuse, what you'd cut, and your honest weakest point. (Some of this is intentionally underspecified — make calls and state your assumptions.)
3. One line: what's something in AI/agents/GTM you started using in the last 2–3 weeks, and why?

Send your link(s) to the recruiter. If it's a fit, you'll do one live session — expect us to be direct, to change a requirement on you mid-conversation, to have you build something small live, and to push hard on your thinking. (If you're running your own thing and a recorded video isn't your style, tell the recruiter — we'll do a live equivalent or jump to a short paid build instead.)