Every candidate who reaches your shortlist on Iber has already done the work you would otherwise spend three rounds discovering.
What a candidate arrives with
A verified build. A demo of something they made, linked to a GitHub repository we confirmed is under their own account. Not a fork they never touched.
A spoken interview about that build. Recorded, proctored, and graded on reasoning, specificity, ownership and technical depth. The questions come partly from a role bank and partly from their own projects, so it is very hard to prepare a script for.
A signed Iber Card. A public credential page with the evidence behind it — builds, results, trust tier — that you can verify rather than take on trust.
The pipeline
You choose how deep the assessment goes when you post the role:
- Junior — one AI screening interview
- Mid-level — screening, then a deeper professional interview, then a browser coding test
- Senior — the above, plus a 60-minute challenge on the candidate's own repository, plus a final extreme coding round
Every stage runs in the candidate's browser. Nothing needs scheduling and nobody on your side sits in a first-round call to find out someone cannot explain their own project.
You can also switch off rounds that do not fit the role — a business analyst does not sit a coding test.
What you see
Ranked applicants with per-stage scores, the full transcripts, the recordings, and a plain-language integrity record from the proctoring.
That last part matters, so we will be precise about it: the proctoring produces observations, not verdicts. Timestamped notes for a human to read — a phone visible, a second person in frame, focus lost. It does not eliminate anyone by itself, and we tell you when a signal is a coarse estimate rather than a fact. You make the call.
Anti-gaming
The obvious attack on a platform like this is submitting work you did not do. So:
- Repo ownership is checked against the candidate's own GitHub account
- The interview asks about their code, and follow-ups react to what they actually said, live
- Senior roles include a timed challenge on their own repository, followed by a spoken exam about the exact diff they just wrote
- Identity is matched against an enrolled selfie before a take starts
- Any attempt to instruct the AI evaluator through a transcript or a README is flagged to you and never moves the score
None of it is perfect. All of it is shown to you as evidence with its confidence attached.
Pricing
Builders are free, permanently. Companies pay to hire — we are working with early partners on pricing while we learn what a placement is worth to you. If you want to be one of them, get in touch.
Getting started
Email shiva@joiniber.com with the role you are hiring for. We will set it up with you.