A good candidate assessment report for an engineering hire fits on one page and answers four questions: can this person do the work, how sure are we, where are they weak, and what should the next interviewer ask. Below is a sample report for a fictional senior backend candidate, so you can see the shape. Use it as a template for your own reports, or as a yardstick for any vendor that sends you one.
Most reports we see from agencies fail the second question. They say "strong candidate" and never say how strong, or compared to whom.
Key Takeaways
- A useful candidate assessment report shows scores, the evidence behind them, and a comparison point. A score with no comparison is just a number.
- Separate what was tested (code, written answers, interview) from what was not tested (team fit, domain depth). Say both.
- Put the risks on the first page, with a concrete question to ask in the next interview.
- The report informs the decision. A person still makes it.
Sample Candidate Assessment Report
Illustration only. The candidate, scores and comments below are invented to show the format.
Candidate: ENG-0000 (anonymized) | Role: Senior Backend Engineer (Python, PostgreSQL) | Assessed: October 2026
| Section | Result |
|---|---|
| Coding assessment | 82 / 100. Hidden tests passed: 14 of 16. Failed tests: pagination edge case, concurrent write. |
| Written technical answers (4) | 3.4 out of 5 across four criteria; clear on tradeoffs, thin on observability |
| Comparison | Score next to two anonymous bench engineers assessed the same way this month: 3.1 and 3.6 |
| Suggested next step | Probe retries and monitoring in a live system design round |
Written-answer breakdown (out of 5)
| Criterion | Score | Evidence |
|---|---|---|
| Technical depth | 3.6 | Explained PostgreSQL row locking correctly and when to use SELECT FOR UPDATE |
| Problem solving | 3.5 | Broke the rate-limiter question into parts before writing anything |
| Communication | 3.2 | Clear, but long answers; needed one prompt to get to the point |
| Practical judgment | 3.3 | Chose a queue over a cron job and said why; missed retry idempotency |
Strengths. Solid data modeling. Writes tests without being asked. Comfortable reading unfamiliar code.
Risks. (1) The concurrent-write test failed, and the interview answer on idempotent retries was vague. For a payments-adjacent role, that matters. (2) Little evidence on production monitoring.
Ask next. "Walk me through how you would make this webhook handler safe to retry." And: "Tell me about the last incident you debugged in production. What did you look at first?"
Not assessed. Team fit, domain knowledge of your product, references.
That's it. One page, and a hiring manager can read it in three minutes.
Why the Comparison Line Matters
A 3.4 means nothing on its own. Is that good? On some interview scales a 3.4 is average. On ours, nobody has ever scored above 4.2, so a 3.5 is strong.
The comparison line fixes that. It puts the candidate's score next to engineers who took the same assessment under the same rules. In the reports Ruzora sends, that shows up as the candidate's number next to two anonymous engineers from our bench, assessed identically, that month. It appears when we have enough engineers assessed on the same rubric to make the comparison fair; coding scores clump together, so that part is sometimes shown only as a band or left out. A comparison that is not fair is worse than no comparison.
What Belongs in a Candidate Assessment Report Sample You Can Trust
When a vendor or a recruiter sends you a report, check for these:
- The actual test. What was the candidate asked to do, for how long, and was it timed?
- Evidence under each score. A quote, a failed test name, a specific answer. "Great communicator" without evidence is an opinion.
- The pass bar. Ours is public: a graded coding assessment at 70 out of 100 or better and an AI-led technical interview at 3.0 out of 5 or better, listed on our engineers page. Ask any vendor for theirs.
- What was not tested. Honest reports say it. Reports that claim to have measured everything are selling.
- Who decides. The report should end with a recommendation for the next step, never a hire decision.
A Concrete Version
You found a candidate yourself, through a referral. Strong resume, good first call, but your team has no senior backend engineer to run the technical round.
With Ruzora's free candidate check, you send the candidate a link. They spend about 25 minutes on a coding challenge in a browser sandbox plus four written answers. You get a scorecard within a day of them finishing. We never contact the candidate, and answers are deleted 90 days after the report. You can check one candidate at a time, up to three per company per month.
If your report looks like the sample above (coding score, written answers scored on four criteria, a comparison line), you know what to probe in your final round: retries and production monitoring. If you want a senior engineer to run that round live, the second opinion is a 45-minute interview with me and a same-day written verdict, $299 per candidate, currently arranged by call. Score that final round with an interview scorecard so the debrief stays on evidence.
The Honest Counterpoint
A 25-minute assessment is a screen. It tells you whether someone clears a bar on code and written reasoning. It won't tell you whether they will thrive on your team, handle your on-call, or get along with your lead. A strong report can still precede a bad hire, and a borderline one can precede a great one.
Scores also invite false precision. The difference between 3.3 and 3.4 is noise. Read bands, not decimals, and always read the evidence before the number. If you find yourself rejecting someone over a tenth of a point, the report has started making your decisions for you.
Frequently Asked Questions
What should a candidate assessment report include?
What was tested, the scores with evidence for each one, a fair comparison point, strengths, risks, the questions to ask next, and what was not assessed. One page is enough.
How do I read a technical interview score?
Look at the scale's real distribution, not its maximum. If nobody scores above 4.2 out of 5, then 3.5 is strong. Ask the vendor what a typical passing score looks like.
Can I get a candidate assessed without hiring through an agency?
Yes. Ruzora's free check assesses a candidate you already found, with no obligation to hire through us. For a deeper look, read how to verify a senior engineer yourself.
The Bottom Line
A candidate assessment report earns its keep when it shows evidence, a fair comparison and the risks to probe next. Use the sample above as your template, and hold every vendor's report to it.
Check a candidate for free or see engineers who already passed.
Roberto Espinoza is CEO of Ruzora, which helps US startups hire pre-vetted senior LATAM engineers, with a vetted shortlist in 72 hours. See available engineers.
