Test the skill,
not the resume.
A sandboxed, in-browser assessment for automation engineers. Candidates write and run real Selenium and REST API tests against a live app — and you get an objectively graded, cheat-resistant result straight in your pipeline.
You can't interview your way to a real test engineer
Automation-testing roles are notoriously hard to screen. Talking about Selenium is not the same as writing a suite that catches the bug in production.
Resumes claim skills they can't prove
"5 years of Selenium" tells you nothing about whether someone writes tests that actually catch bugs. The keyword is on the page; the skill may not be.
Take-home tests get gamed
Unproctored take-homes are copy-pasted from GitHub or generated by an LLM in minutes. You end up interviewing the internet, not the candidate.
Manual grading doesn't scale
Reading test code by hand is slow, subjective, and inconsistent between reviewers — so strong candidates slip through and weak ones advance.
A real coding environment, graded automatically
No local setup for the candidate, no manual grading for you.
01
Recruiter picks a problem
Choose a problem by app, skill, language, and seniority — junior element drills through senior REST API design — and send the candidate a signed link. No scheduling, no proctor on a call.
02
Candidate writes and runs tests
The candidate opens a Monaco IDE with a starter project next to a live preview of a real app under test, writes their suite, and hits Run — real execution in a per-session sandbox, not a syntax check.
03
We grade it objectively
On submit, correctness is scored by mutation testing or scenario coverage, code quality by static analysis, and an advisory integrity band summarises edit telemetry. Same rubric for every candidate.
04
A report lands in your pipeline
A decision-ready report is delivered back to Hiearlly over an HMAC-signed webhook, so the candidate advances in the same flow as sourcing, ranking, and screening.
A score you can actually trust
Four independent signals combine into one defensible result — every candidate on the same rubric.
Mutation testing
We seed bugs into the app and check whether the candidate's suite catches them. A test that passes but never fails on a broken build isn't a real test — mutation scoring finds that out.
Scenario coverage
Where a mutant bank isn't available, we score how much of the intended behaviour the suite actually exercises — flake-resistant, with repeated golden and candidate runs.
Static analysis
Tree-sitter parses the submission for code quality and test design — structure, assertions, and maintainability — not just whether it turned green.
Integrity band
Edit telemetry flags paste and AI-assist signals as an advisory band for the recruiter. It adds context — it never auto-rejects a candidate.
Problems for every level
Each problem pairs a real app with a skill, language, and seniority. Adding your own is a matter of seeding a template.
Web UI automation via Selenium and REST API testing today · Appium / mobile on the roadmap.
Part of the same hiring flow
A skills assessment is only useful if the result reaches the people making the call. When grading finishes, Hiearlly Assessments pushes a structured report back over an HMAC-signed webhook — so the outcome lands in the same pipeline as sourcing, comparative ranking, and agent-led screening.
Source and rank with the Re-ranking Engine, run the first interview automatically, verify hands-on skill with an assessment, and advance the candidate — all from one conversation.
Assessment vs the take-home
Why real execution beats a trusted zip file.
See Hiearlly Assessments live
Bring the roles you're hiring for and we'll walk you through issuing an assessment, the candidate experience, and the graded report — end to end.
Book a demoNo commitment · Free walkthrough · 30 minutes