Hire the people who know when AI is wrong.
Candidates review work an AI produced, with flaws we planted on purpose. About a minute after they submit, you see which flaws they found and which they missed.
This sweep should settle pending stock once, then expire each stale hold.
One of these lines is wrong. Click it to circle it.
pendingStockSkus is built straight from the lines without deduplicating or sorting, so duplicate SKUs could be listed twice and cause a double stock application. It should be wrapped in a Set and sorted before the CAS write.
Dismissed by the candidate. placeHold already rejects duplicate SKUs in a reservation, so the list cannot repeat.
Every exercise comes with its answer key.
Pick an exercise or generate one for the role, then send the link. A draft score is ready about a minute after the candidate submits.
Everyone on the same exercise is marked against the same key, and a person checks each score and signs it.
| Planted flaw | Outcome | Quality |
|---|---|---|
| security | Caught alone | 2 |
| logic bug | Caught alone | 2 |
| hollow test | Caught with the assistant | 1 |
| performance | Missed | |
| overengineering | Caught alone | 1 |
| false claim | Caught alone | 2 |
Finalized 22 Oct by an interviewer, with a recommendation to hire.
Run the rest of hiring from the same record.
Stages, stage emails, interviews and scores sit on one profile. Rankr drafts each email from your template, and a person sends it.
Already on an applicant tracking system? The API and signed webhooks cover the same steps.
- Applied14 Oct
- Screening16 Oct
- Interviewing21 OctCurrent stage
- Offer
- Hired
- Document critiqueFinalized at 7 / 9
- Screening emailSent 16 Oct with the scheduling link
- Case study with the finance leadInvitation sent 21 Oct
- Team conversation with the head of operationsInvitation not sent yet
Each role gets an exercise from its own work.
- EngineersPR critique, Challenge the AI
- Senior engineers and architectsDesign review
- Data scientists and analystsVerify the claims
- Product, finance, operations and peopleDocument critique
- Sales and marketingRank the answers
- Any roleRank the answers, Rapid fire
The prices are on the page.
It costs less than every other assessment platform we know of. Pay monthly, or pay yearly and get two months free.
- Free$0 a month
5 assessments and 2 generated items a month, 3 members.
- Starter$29 a month
25 assessments and 10 generated items a month, 5 members.
- Team$69 a month
100 assessments and 40 generated items a month, 20 members.
- Business$199 a month
400 assessments and 150 generated items a month, unlimited members.
| Free | Starter | Team | Business | |
|---|---|---|---|---|
| A month | $0 | $29 | $69 | $199 |
| Assessments a month | 5 | 25 | 100 | 400 |
| Generated items a month | 2 | 10 | 40 | 150 |
| Members | 3 | 5 | 20 | Unlimited |
| Start free | Choose Starter | Choose Team | Choose Business |
You decide what Rankr keeps about candidates.
- Everything stays inside your organization, and no model sees a candidate's CV while scoring. How data is kept apart
- You can export or erase a candidate at any time, and closed candidates are erased after the period you set. Privacy policy
- SOC 2 is on the roadmap, and the security page says what is in place today.
Run your first assessment this afternoon.
Rankr started as the exercise we gave our own candidates at Gearlay, after take-home tests stopped telling us anything. There are no customer logos here yet, and we'd rather be judged on the exercise.
For the Business plan or more than one organization, talk to us.