I’d reduce filtered-search latency with pagination and cached queries. What else should we consider?
Technical interviews for the AI-native engineering era.
Meritoso is an AI-native assessment platform for realistic SWE work-sample interviews — where AI use is allowed, transparent, and evaluated.
Candidate task: Improve search performance
Good start. Other options include selecting fewer fields, adding filter indexes, or debouncing requests.
Let’s use pagination plus filter indexes. That should improve first-load time without taking on cache invalidation. Please implement it with a filtered-pagination test
Sounds reasonable — I’ll make the change.
The old technical screen is losing trust.
AI assistance is now normal in software work, but many hiring loops still treat it as invisible cheating or try to solve it with detection-first controls.
LeetCode-style tasks reward memorization.
Teams need evidence of debugging, testing, code review, and tradeoff judgment.
AI bans are brittle.
Hidden AI use, overlays, and second-screen help make simple pass/fail screens noisy.
Take-homes create candidate friction.
Unbounded assignments can feel like unpaid labor and still leave authorship unclear.
Turn AI use into hiring signal — not hidden risk.
Meritoso gives candidates a clear AI-use policy and helps hiring teams evaluate the judgment around the work: how candidates steer tools, verify outputs, test decisions, and explain tradeoffs.
Problem framing
AI steering
Debugging
Testing
Tradeoffs
Code ownership
How a Meritoso interview works
A focused 60-90 minute work-sample assessment built to show how candidates reason, verify, and improve real code.
Choose a realistic repo task.
Debug, extend, review, or triage code that resembles day-to-day engineering work.
Give candidates a clear AI-use policy.
Candidates know what is allowed, what should be disclosed, and what will be evaluated.
Review the diff and evidence.
Look at the solution, tests, reflection, AI-use summary, and follow-up prompts.
Review with AI insights and a human rubric.
Hiring teams evaluate problem framing, tests, and tradeoffs while keeping decisions with humans.
Candidate task: Review an AI-generated PR
Add checkout retry support
3 files changed
Backend
Feature
42
const existing = await db.payment.findOne(q);43
if (existing) return existing;44
+await db.payment.create({45
+ orderId,46
+ idempotencyKey,47
+ status: "PENDING"48
+});Candidate task findings
Missed risk
DB-level duplicate guard- Notes retry-path test coverage
- Asks about concurrent retries
Captured signal
Strong test instincts; missed DB guard.Built for teams hiring in the AI-native engineering era.
Meritoso is for engineering-heavy companies where engineering, technical recruiting, and talent operations each need clearer answers from the same work-sample evidence.
Engineering managers and leaders
See whether candidates can debug, verify, and explain AI-assisted engineering work.
Technical recruiting
Set clear AI-use expectations and give candidates a fairer, more realistic alternative to brittle screens.
Talent operations
Standardize the evidence teams use for calibration, process visibility, and assessment quality.
Questions buyers are asking
If AI is allowed, what are we evaluating?
How candidates work with AI: framing the problem, steering the assistant, verifying output, writing tests, debugging failures, explaining tradeoffs, and owning the final code.
Is this proctoring?
No. Meritoso is not surveillance-first. The goal is to make AI use transparent and reviewable, while treating identity or proxy-candidate checks as a separate integrity problem.
Is Meritoso making automated hiring decisions?
No. Meritoso supports human review with structured evidence, rubrics, and follow-up prompts. Hiring decisions stay with the team.
Is this just another take-home?
No. The goal is a bounded 60-90 minute, job-relevant work sample with clear expectations and a focused follow-up discussion.
How is this different from a coding test?
Meritoso focuses on realistic repo work, transparent AI-use rules, process evidence, and human judgment — not puzzle completion or final-code correctness alone.
What do teams get from the interview?
A structured view of the candidate’s solution, tests, AI-use summary, reflection, and rubric signals for follow-up and calibration.
Help shape better AI-era technical interviews.
We’re speaking with engineering and talent leaders adapting software interviews for a world where AI is part of real work. If this is on your roadmap, schedule a short call and let’s compare notes.
Schedule a call