FOR HIRING LEADERS
Hiring, in production.
We’re building SourceWork: take-home exercises that run in production. Candidates get a real repo, a CI pipeline and live traffic to handle, and use AI as they would at work. Your rubric is applied to what they ship, so reviewers know where to look and what to ask at interview.
Survey respondents get early access and the results report when the survey closes.
Scoring guidance per rubric criterion, with evidence.
No scores. Where a reviewer should read closely.
Interview questions from their own code. Can they explain what they built?
Spend valuable hiring time on the things that matter. Automate the rest.
WHAT WE’RE BUILDING
An exercise that behaves like the job.
Puzzles and whiteboards measure how someone interviews. Real work, under real conditions, shows how they engineer.
- 01
Set the exercise
Start from a library of production simulations that we write and maintain, or build your own with your traffic shape and failure modes.
- 02
Candidates ship
Each candidate gets a repo and a pipeline. Their code, in any language and written with the AI tools they use at work, deploys into a simulated environment and takes real requests.
- 03
Review with signal
Your rubric is applied to the code, the pipeline, and how the service held up. Reviewers get everything they need to make the right call, quickly, including questions to ask about the candidate’s own code.
BUILT FOR BOTH SIDES OF THE DESK
Better signal for teams. A fairer test for engineers.
FOR ENGINEERS
Show what you can do in a real environment, in the language you work in.
Help us build it right.
We’re talking to hiring leaders and engineers before we ship. Five minutes of your experience decides what we build first.