Forward Deployed Engineer

About us

ironco is an AI growth agency. We run growth for consumer and B2B brands, starting with TikTok Shop: affiliates, creator content, samples and the reporting that comes from them. We're an AI-first team. Agents do the work, and people run the agents.

The role

You own the agents our pods run. You build them, prove they work, and run each new one yourself on a live pod until the pod can run it without you. Small fixes ship the same day. Bigger builds go to the product engineers with a spec. When a pod stops using an agent, that's yours to fix. This is a hands-on seat: there's no team to lead, and the work is about half building and half operating. AI writes most of your code, and you own every line of it.

What you'll do

• Build and prove the agents: take a workflow the product team has scoped, build the agent, and write the checks that show it does the job. You own its numbers: which pods use it, the share of cases it handles, and the share sent with no human edit.
• Run each beta yourself on a live pod: approve the samples, reply to the messages and watch the automations. Fix what breaks the same day, and write the bigger builds up so an engineer can start without asking you anything.
• Hand it over clean: when a pod runs the agent without you and the exceptions are written down, operations takes it and you start the next one. Show the hours it took off a human, before and after.
• By day 90: pods run agents you built every day without you, and you can show the hours they save.

Who you are

• You've built something with AI that other people use, and you can demo it: an agent, an automation or an internal tool. Built, not just used.
• At least two years shipping software, in any stack. You ship with AI every day and check what it writes, with tests or by running it on real data before you trust it.
• You'd rather sit with the user than guess: you watch where the agent fails, then fix the instruction or the code. You care whether it gets used, not only whether it runs.
• Clear written English, and happy to work afternoons and evenings for US overlap.
• What you get: you build and run AI agents on real client work from your first week, with room to grow into a senior engineering or product role.

Not for you if

• You want to manage a team or set the roadmap.
• You'd rather hand off a spec than run a half-finished tool yourself on a live account.

How hiring works

The screener case is required for everyone. Each stage is a pass or a stop. The proof is in the work: can you do this job well, and will you still enjoy it a year from now? Most candidates stop before the team stage. If it ends at any stage, we'll tell you where you fell short.

1. 1

   Screen

   A 15-minute call with our recruiter, and a 5-minute recording of how you work with AI. If both pass, the screener case.

2. 2

   Interview

   15 minutes with Jeremy Carr, ironco's founder and CEO, sharing your screen. If it passes, a longer project of about 2 hours.

3. 3

   Team

   A working session on your project with a few people from the team, then one or two final interviews.

The 5-minute recording. Your screen and your voice, in one take, with Loom or any screen recorder.

1. Your setup (1 minute): the AI tools you use every day, and the instructions behind them.
2. One live change (3 minutes): in something you built or use, change one instruction, rerun it, and show how you check the result is right.
3. One miss (1 minute): a time the AI got it wrong, how you caught it, and what you changed so it doesn't happen again.

Preparing is fine. Talk us through it the way you'd show a teammate.