For your first hires.
Up to 50 resumes read and 2 interviews run, each month.
- Every resume read and scored
- One ranked shortlist per role
- Nothing counts until it happens
Built for recruiters.
We read every resume, run the coding interview, and hand you one shortlist.
Indeed
Glassdoor1 2 3 4 5def solution(records): seen = {} for name, ok in records: seen.setdefault(name, set()).add(ok) return sorted(seen)
But how do you find the right technical talent in a sea of resumes that all look the same?
Custom questions. Real proctoring. Honest results.
Your CI harness records the outcome of every test across repeated runs. You are given a list of run records, each a pair [test_name, passed] where passed is a boolean. A test is considered flaky if, across all its recorded runs, it has at least one pass AND at least one failure. Return the list of flaky test names sorted alphabetically. Tests that always pass or always fail are not flaky.
Input: [["login",true],["login",false],["search",true],
["search",true],["checkout",false]]
Output: ["login"]login has both a pass and a fail; search always passes; checkout always fails.
Input: [["a",false],["a",false],["b",true],["b",false]]
Output: ["b"]a always fails so is stable; b flaps between pass and fail.
0 ≤ number of records ≤ 20000
test_name is a lowercase string with at least one character
1def solution(records): 2 # records: list of [test_name, passed] 3 pass
248 applicants, 5 recommended for your review.
Recrewit runs the whole technical search. It reads every resume, runs every technical interview, and scores it all. No more guessing whether a candidate can actually do the job.
Every resume that comes in is scored against your job description.
Technical questions are generated from your job description and the interview is proctored start to finish.
Camera and screen stay on with detection running and every minute recorded for your review.
Every score is checked twice and lands with the reasoning behind it.
You take it from here and choose who moves forward.
You’re only ever counting two things: resumes we read and interviews we run. The shortlist is always included.
Up to 50 resumes read and 2 interviews run, each month.
Up to 250 resumes read and 10 interviews run, each month.
Reads and interviews sized to your busiest season.
I used to read two hundred resumes to find five worth a call. Now those five are waiting for me on Monday morning.
I am not an engineer. I can still defend every shortlist I hand a hiring manager, line by line.
112 assessments on our first deployment at an insurance client, and 63 roles filled. Six months on, Recrewit is part of how we run every search.
No. Recrewit scores and ranks. A recruiter decides. Every score carries plain language reasoning and an override, so nothing is rejected by a number you cannot see or argue with.
No. That is the point. Recrewit reads the code and tells you what it found in plain English, so you can defend a shortlist to a hiring manager line by line without writing a line yourself.
Protected signals are redacted before anything reaches the model: names, pronouns, graduation years, addresses. AI likelihood is reported as its own flag and never folded into the fit score, so a candidate is never marked down twice for one thing.
No. Recrewit does the screening your ATS was never built to do, then hands back a ranked, reasoned shortlist. Your ATS stays the system of record.
Hidden test cases never leave the server, so there is nothing to farm. Typing behaviour, tab switches and the session recording are all reported to you, as signal you weigh, not a verdict we pass.
Our team sets it up with you and stays on through your first role. Most teams post the same week. There is nothing to install, and candidates run the interview in a browser.
One email a month, from the team reading every resume.
One email a month. Unsubscribe any time. See our privacy rules.