-
-
TYPE OF WORK
Full Time
-
-
SALARY
$2,500-$3,500 (with performance)
-
-
HOURS PER WEEK
40
-
-
DATE UPDATED
Jul 7, 2026
We're an AI-native B2B outbound company. We build the systems that source leads, run campaigns, and handle replies at scale. You'd work across our internal engineering: building and maintaining sourcing modules, developing internal SaaS features, keeping systems running, and researching new ways to use Claude Code to build lead lists and ship faster.
This role is remote. Compensation is a competitive base of $2,00+/month plus uncapped performance bonuses. Bonuses are tied to what you actually produce: number of modules and features shipped per month, bugs fixed, and reliability metrics on the systems you own. A strong performer clears $3,500+/month all in, with no ceiling above that. The engineers who ship fast and ship clean earn the most.
This is a demanding environment and a rewarding one. If you want a quiet 9-to-5 where nobody notices what you do, this isn't it. If you want to build real systems, own them end to end, and get paid on what you produce, keep reading.
What you'll do
Build and maintain sourcing modules, develop and maintain internal SaaS features, and research new ways to leverage Claude Code for building lead lists and speeding up development. You'll work across the stack with tools like Clay, HeyReach, Smartlead, Supabase, and n8n.
How you're paid
Base plus uncapped performance bonuses. The variable is measured, not vibes. We track:
Volume shipped: modules and features delivered per month
Bugs fixed: issues you close across our systems
Reliability: uptime and error rates on the systems you own
Hit those and you go past $3,500/month. There's no ceiling. Coast and you sit at base.
How the application works
We don't read CVs first. We watch you build.
You'll be given a test task: build one working sourcing module. A sourcing module is a small program that pulls a list of companies or contacts from a source of your choosing (a directory, a job board, an API, a scrape, whatever you like), cleans and standardises that data into a consistent format, and sends it to an endpoint we give you. It's the piece that turns a raw source of leads into clean, structured, ready-to-use data.
This is a self-contained project that should take no more than about three hours. You're expected to use AI tools like Claude Code or Codex to build it fast. That's how we work here.
We run this on a group interview call. Everyone gets sent the SOP and the specs, and you're expected to work out what needs to be done yourself. We're hiring for initiative. Nobody here gets handheld, and the test is partly whether you can take a spec and run with it. The brief covers the inputs, the outputs, and what we mean by useful internally, so you'll have everything you need to build without asking.
We judge every submission on three things:
Reliability: does it work every time
Error handling: what happens when the source is messy, incomplete, or breaks
Usefulness: is this something we'd actually run internally
We don't care how clean the code is, only that it hits those three.
We're picking one engineer for this role. The best module wins it, with the performance bonus on top.
To apply
Record a short custom Loom telling us why you're a fit and what you'd bring. In that Loom, say the word "pipeline" out loud so we know you read this the whole way through. Applications without the word get binned.
Keep the Loom under 5 minutes. We care about how you think, not production value.
Then send it along with anything you think earns you a spot.
