Assessment Track · How It Works
A hiring test built like the job not a quiz about it.
Every candidate works one continuous scenario: a fictional client, a real problem, and about three hours to run the engagement start to finish. No proctor, no whiteboard. What comes out the other side is a full evidence report, not a pass/fail number.
What The Candidate Actually Does
Five stages, one client, no do-overs.
Each stage produces something a real engagement would produce — a discovery summary, a plan, a working fix, a handoff document. The candidate can't skip ahead, and once a stage is submitted, it's locked.
Discovery conversation
The candidate is dropped into a call with a busy, non-technical stakeholder who has one vague complaint and no patience for jargon. They have to find the real problem, its cost to the business, who else is affected, and what's actually off-limits — before they know anything about the fix.
Solution plan
Before touching anything, the candidate commits in writing: their leading hypothesis and the first moves they'd make to test it. This is where a guess looks like a guess — on paper, before it ever reaches the system.
Hands-on technical work
A real, broken environment — or a live one, actively misbehaving. The candidate fixes it the way they'd fix it on a client's system on day one: same tools, same mess, same clock.
Prevention & handoff plan
With the fix in place, they write what a client executive would actually receive: what broke, what changed, what's still unknown, and what happens next. Written for someone who wasn't in the room.
Presentation
A short recorded walkthrough, delivered the way they'd present it to the client's leadership — what was built, what the risks are, what to do next.
Why It Can't Be Rehearsed
Every stage depends on the one before it.
Take-home tests get shared, memorized, and solved by someone else. This doesn't work that way. The specific facts a candidate needs to solve the technical problem are the same facts they had to actually notice back in the discovery conversation — not a separate, unrelated puzzle bolted on afterward.
A candidate who guesses or gets lucky at one stage runs out of runway at the next.
How It's Graded
Two layers: what happened, and how they got there.
Neither layer is a black box. Every score in the final report links back to the actual transcript, code, or document it came from.
Layer 1
Did it actually work?
The fix runs against the real system it was meant to fix. This part isn't an opinion — the service either responds correctly or it doesn't, and that result is recorded exactly as it happened.
Layer 2
How did they get there?
Every written response and the actual code change is reviewed for reasoning and judgment — the way a senior colleague would review the work, not a search for magic words.
Sample evidence report
Illustrative — not a real candidateWant to see the full report format?
Create an account and send yourself a sample assessment.