Founding Customer Engineer
Seattle · Full-time
You take a customer from "we signed up" to "our AI teammate is live in a client channel and the client just replied." Then you turn what you learned into the playbook that gets the next ten customers there faster.
Who we are
We fight for the small companies.
Five million people in America make a living by putting their judgment to work inside other people's businesses: agencies, consultants, bookkeepers, recruiters. They are already trusted and already inside millions of companies. Nobody has built AI for how they work.
The big AI players build for the enterprise: one agent inside one company, and the expertise ends up owned by the platform. We believe expertise should belong to the people who built it. An agency works differently. One expert serves many clients at once, and no client may ever see another client's world. So we build AI teammates for that: they live in the agency's own tools, learn by watching the agency work, and go to work for each client under the agency's name.
We were our own first customer. AJ grew an agency to twenty-four people, Dan ran a software studio at three million a year, and we built Ergoly inside the agency we ran together. Today agencies run real client work through it, and every week another consultant tells us they are teaching themselves to code to stay ahead. Those are our people. The big players got the enterprise. We are here for the five million.
The team is small. Nothing is someone else's department, there is no layer between building something and shipping it, and what you build reaches real agencies and their clients within days.
The role
Two workstreams, running in parallel from the start.
The deployments
A customer signs on Monday. By Tuesday you have read their tools, their client list, and their Slack, and you can say: here are the three AI teammates worth building, this one first, and it goes live Thursday. Then you make that happen — set it up, connect it, ship it — and you live in the shared channel with them, where most of what you write is three lines long.
The playbook
Customer four teaches you something. Customer five gets it on day one. You write down every setup step, learn every place customers get stuck, and turn every fix into a template — until a new customer goes live on a playbook anyone can run, including our own AI teammates.
Who this is for
You like hard problems? Talking to customers until you understand theirs? Driving when there is no structure around you yet?
Then read the next six lines slowly. They are about you.
- You get a kick out of understanding someone else's world fast, and drawing it back to them better than they explained it. You can follow an API doc, a Zapier flow, or a database schema and build the small thing that connects them.
- You want to talk to customers. Real ones, every day, in their Slack, about what is broken and what they wish existed.
- You have built something with an LLM and felt the moment it worked. You want that feeling again, with someone else's business riding on it.
- "Find out why customers go quiet between day 3 and day 10" is the assignment you would pick for yourself.
- You drive when there is no structure yet. You spot what is broken, fix it, and write it down so it stays fixed.
- You know service businesses from the inside: what a retainer, a client channel, and a burned-out account lead look like.
If you have done DevRel, solutions engineering, or forward-deployed work, or you were the technical person at an agency who made everything fit together, you already know this job.
What a good week looks like
Most of this job happens in a shared channel.
A customer asks for an AI teammate. You answer with how it is built, and it is live by Thursday. That week reads like this.
# ergoly-northwind · shared channel
Maya · NorthwindTue 9:14
We want an AI teammate that drafts the weekly Bluebird status update. It needs Linear for what shipped, HubSpot for the open deals, and it has to sound like Priya wrote it. Where does all of that live?
YouTue 9:18
Three pieces. Linear and HubSpot connect as tools on the AI teammate. Priya's last six updates go into the Bluebird context so it learns her voice. The "lead with the win" rule goes into your house rules so every client gets it. Wiring it now — first draft lands in the Bluebird channel Thursday morning.
YouThu 16:02
Did it work?
Maya · NorthwindThu 16:09
Yes. Priya sent it with one edit. Can we do the same for the other four?
The actual problem
You are the reason the thing works, for every customer we sign.
If that sentence makes you lean forward, keep reading.
Ergoly has no repeatable path from signed to live yet. The founders walk every customer through it by hand, and each one needs something new. You find the path, and you prove it the only way that counts: a customer's client typing a reply to an AI teammate you put there.
For the right person, that is the most interesting job on the team.
What is in it for you
- Getting customers live is yours from day one: the number the whole company watches, the playbook, and the team that comes after you. How a customer goes live is your call.
- Pre-seed equity. The ground floor of something that does not exist anywhere else yet.
- A product early enough for your fingerprints to be all over it, and customers who tell you what they think within the hour.
- The founders and the customers, every day, in the same channels.
The numbers
- Work trial
- One week, paid, full-time
- After the trial
- Competitive full-time comp + pre-seed equity
- Location
- Seattle
- Reports to
- AJ, cofounder
One week, paid, on real customers in real channels. It is the same job on both sides of the trial, and by the end of it we both know.
How to apply
Three questions.
No cover letter, no take-home, no six rounds. Three questions, read by the two of us. You hear back from AJ or Dan either way.
You can also apply from a terminal: ssh apply@join.ergoly.ai asks the same questions and reaches the same two people.