Forward-Deployed AI Engineer
Indexed description
READ THIS FIRST: TO BE CONSIDERED, YOU MUST SUBMIT A 5-MINUTE LOOM VIDEO
Applications without a Loom will not be reviewed. Record yourself (camera on, in English) walking through one real thing you built with AI: the business problem, the architecture and two or three key decisions, the hardest technical problem and exactly how you solved it, and what you would do differently today.
Email the link in your application to [email protected]. The Loom video should be 5ive minutes, not fifteen. How you communicate in that video is the first thing we evaluate, because it is a large part of the job.
THE SHORT VERSION (TL:DR)
Most engineering jobs hand you a ticket. This one hands you a client.
You will build production AI systems at a speed that was not possible three years ago, using Claude Code, Cursor and whatever agentic coding terminal wins next month. You will own the scope, the timeline and the budget. And you will sit on the weekly call with the COO or the plant manager, explain in plain English what you shipped and why it matters, and keep that client confident enough to stay another month. Our contracts are month-to-month with no minimums. We earn every client every month. So will you.
This is a forward-deployed role. Client-facing is not a side duty. It is the job.
WHO WE ARE
Zaigo Labs builds custom AI agents, automations and bespoke software for mid-market and private-equity-backed companies in the United States: healthcare, insurance, manufacturing, construction, waste services, financial services, even college football. Three years old, profitable, bootstrapped, about 25 people. Working systems in production, or nothing.
Recent work, described loosely: an AP automation stack that reads and routes tens of thousands of vendor invoices; a renewals scrubber and submission agent for a specialty insurer; consent, pricing and invoice validation for a multi-site hospice operator; inventory intelligence and cash forecasting for a manufacturer; a firmwide AI knowledge layer for a PE fund; a recruiting and game-planning system for a Power 4 football program.
Each of those began as an operator's vague complaint and ended as software people rely on daily. One engineer owned that entire arc. We are hiring more of them in Istanbul.
THE THREE THINGS THAT MATTER
1. World-class at AI-native building, and relentless about getting it done.
You live in the coding terminal. You architect the system and let the agents write most of the code, then you read it, test it and ship it. When something blocks you, you find another way. You do not wait to be unblocked. You treat every client system like you own it, because in practice you do.
2. You run the engagement, not just the code.
On time, on spec, on budget. You break the roadmap into checkpoints, hit them, and already know what the client needs next before they do. Part engineer, part product manager, part program manager. If you only want to code, this is the wrong posting.
3. You are excellent with clients.
Fluent, polished English. You present well, listen better, and can unpack a complex build for a non-technical executive without dumbing it down. You enjoy the relationship side, you sense when a client is nervous, and you know how to earn trust and expand the work. The engineers who thrive here are a little bit salespeople and proud of it.
A NORMAL WEEK
Monday, a 30-minute checkpoint with a US client's operations lead: what shipped, what is next, where the roadmap stands. Tuesday through Thursday, heads down building in Claude Code/Codex/Kimi etc: an ingestion pipeline, an agent that reconciles two systems that have never talked to each other, a clean internal UI. Friday, a Loom for the client walking through the new release, and a short note on what you recommend they do next. Every week you are closer to the client than the week before.
WHO THIS IS FOR
- You have shipped real software that real people depend on, before and after the AI coding era
- You adopted Claude Code, Cursor or Codex the week they came out and have not looked back
- You have built with LLM APIs, RAG, vector stores, MCPs and agent frameworks in production, not in a notebook
- You are fluent in our stack or can be in a week: TypeScript, Next.js, Python, Supabase, Postgres, Vercel, plus the usual AWS and GCP
- Your spoken and written English is genuinely strong. You will be judged on it on day one, by the client
- You want to be on the call, not hiding behind the Slack channel
- You can work meaningful overlap with US business hours
WHO THIS IS NOT FOR
- You want a spec handed to you
- You see client calls as an interruption to the real work
- You need a manager to tell you what to do next
- You are uncomfortable being accountable for a deadline and a budget
WHAT YOU GET
Real ownership from week one. Direct exposure to US operators and PE-backed companies. Founders who will put you in front of the client early and back you when you are there. A flat, profitable, no-nonsense company with no layers between you and the work. Compensation that reflects senior-level responsibility. And the kind of experience that makes you the engineer companies fight over: someone who can build, run and sell an AI system end to end.
HOW TO APPLY
1. Submit your 5-minute Loom (see the top of this posting). Include a link to something you built.
2. A working session with one of our senior engineers on a real Zaigo-style problem.
3. A conversation with the founders.
We move quickly and tell you where you stand at every step!
Create a free Caio profile to unlock more results and save your role and location preferences.
Unlock free search