For senior engineers
Work with Outcraft AI.
We work with senior engineers who can ship production software with agentic coding workflows, own ambiguous problems, and stand behind their work when it reaches real users.
Why join
Client work for engineers who do not need hand-holding.
Outcraft is intentionally selective because the client promise depends on it. If you clear the bar, the goal is to put you in front of serious teams with clear problems, not into an endless recruiter funnel.
Serious client work
No recruiter spam or race-to-the-bottom marketplace. We introduce you when the client has a real problem, budget and success criteria.
Flexible engagements
Most work is remote contract delivery, from focused part-time help to deeper monthly engagements.
Founder-reviewed applications
Every engineer is reviewed against the same published standard. The client promise only works if the standard stays high.
Clear expectations
Before a trial starts, you know the scope, repo context, communication rhythm and what the client will judge.
Who fits
Senior first. AI-native second.
Tool fluency matters, but the bar is not prompt cleverness. We care whether you can make good engineering decisions when the work is ambiguous and the model is not always right.
- Evidence of production ownership matters more than years of experience.
- You have owned at least one system end to end: architecture, deployment, failure modes and support.
- You use AI coding tools as part of a disciplined workflow, not as a substitute for engineering judgment.
- You can explain tradeoffs clearly to founders, product leaders and other engineers.
- You are comfortable with remote client work, written updates and scoped trial outcomes.
The Standard
Four stages. Published so you can prepare.
No trick questions, no whiteboard trivia, and no take-home that is secretly free client work. You will know what you are being assessed on.
01
Proof of ownership
Years do not qualify someone. Evidence does. We look for shipped work they can explain clearly: GitHub projects, production AI features, private repo walkthroughs, technical write-ups, or systems they owned end to end.
02
Systems interview
We are not testing whether they know system-design vocabulary. We are testing whether they can take a messy founder problem, shape it into a system that can ship, and clearly explain the tradeoffs and risks.
03
AI-native build
They complete a realistic product task using an agentic workflow: understand the repo, plan the change, implement it, test it, and explain how AI changed their delivery loop. If the task involves LLMs, we also look for evals, cost awareness, and failure paths.
04
Adversarial review
They defend their own code while we attack it. We're looking for engineers who can name every tradeoff they made, not ones who got a good result by accident.
Apply
Send the evidence, not a generic CV.
The strongest application explains a system you owned, what it had to survive, the tradeoffs you made, and how AI changed your delivery loop without lowering the engineering bar.
Include this
- LinkedIn or personal site
- GitHub or code sample
- Best production system you owned
- AI tools you use seriously
- Availability and expected rate
Rates are not hidden. Tell us what you expect, what type of work you want, and how much availability you have.