For hiring teams

600 applied. Five are worth your time. Finding them shouldn't cost your week.

You didn't get into this to read résumés until midnight or pull your best engineers off real work to run first rounds. Paraakh does the screening — fairly, in the browser — and hands you the shortlist worth interviewing.

Create a drive → See it in action

live drive · backend engineer LIVE

600Invited

142In progress

418Submitted

4Flagged

CandidateScoreStatus

PSPriya S.91Shortlisted

LWLin Wei84Shortlisted

DRDiego R.68Below cutoff

You know this feeling

The pile is huge. The signal is tiny.

It's late, the inbox has 300 résumés, and every one of them looks the same on paper.

Your best engineer just spent an afternoon interviewing someone who couldn't write a loop.

You're shortlisting by hand — opening every test, comparing in a spreadsheet, second-guessing.

By the time you've filtered, the one candidate you wanted has taken another offer.

From 600 résumés to your shortlist

The drive runs itself. You just meet the good ones.

Six hundred in. The thirty worth meeting, out.

One link goes to every applicant. They sit it in the browser, the system shortlists on your cut-off, and your seniors interview the handful who earned it — not the whole pile.

See what they can build, not what they typed.

Real coding against hidden tests, across ten languages, with partial credit. A résumé says “proficient in Python.” This shows you whether that's true.

Meet the person before you meet the person.

AI video questions surface communication, confidence, and how someone thinks under a little pressure — so the interview isn't the first time you get a read on them.

No app to install. No candidate lost at the door.

One link, runs in any browser. The strong candidate you wanted doesn't quietly drop off because step one was a download.

Catch the cheating without punishing the honest.

Violations are logged as evidence with timestamps for you to review — a person decides, it's never a silent auto-fail that loses you a good hire over a flaky webcam.

The panel walks in already knowing.

A scorecard with the section breakdown, the skill radar, and the flagged moments — read before the interview, exported straight to your ATS.

What moves

What changes for you.

Your evenings back Auto-shortlist instead of manual review

Senior time protected Engineers interview the right people only

A shortlist you trust Verified skill, not résumé keywords

Fewer ones that got away Move on strong candidates faster

Multi-round hiring pipelines

Turn 500 applicants into your shortlist — automatically.

Chain exams into rounds. Set a cut-off between each one. Add a cooldown if you need candidates to breathe between technical and final. Paraakh advances who qualifies and quietly removes who doesn't.

START 500 applicants One invite link · no signup

ROUND 1

Aptitude & Reasoning

30-min proctored MCQ
Score gate: 60%+ to advance
Auto-grade on submit

→ 300 advance

ROUND 2

Technical screen

Role-specific MCQ + short answer
Cooldown: 24 h between rounds
Score gate: 65%+ to advance

→ 150 advance

ROUND 3

Coding challenge

2 problems · hidden test cases
10 languages · Monaco editor
Auto-shortlist by percentile

→ 20 shortlisted

FINAL Top 20 Ranked, with evidence

Score gatesBlock weak candidates from expensive later rounds.

CooldownsMandatory wait time between rounds — set per step.

Linear or free flowForce a sequence, or let candidates choose order.

Guest invitationsJWT links — no signup, time-bounded, revocable.

Auto-shortlisting engine

Set the criteria. The shortlist writes itself.

Stop opening every result and sorting in a spreadsheet. Combine score, percentile, top-N, and section cut-offs — the engine applies them across the whole cohort in seconds, with a decision and a reason saved per candidate.

RULE 01 Minimum overall score

65%

RULE 02 Percentile cut-off

top 25%

RULE 03 Top N by rank

Top 20 candidates

RULE 04 Section cut-offs

Coding ≥ 60% Aptitude ≥ 50%

LOGIC AND across all enabled rules A candidate must satisfy every active rule to be shortlisted. Each decision is saved with the criteria applied.

shortlisting · applied 118 → 20

418Completed

20Shortlisted

4.8%Shortlist rate

Priya S.91 · top 4%SHORTLISTED

Lin Wei87 · top 9%SHORTLISTED

Arjun M.74 · top 38%ON HOLD

Diego R.61 · below cutREJECTED

Every decision is exportable to CSV or your ATS.

Proctoring layers

Ten layers of integrity. Mix and match per exam.

Strict for a high-stakes certification, lenient for a practice quiz — every layer is configurable, every violation is logged with evidence.

📷 Webcam capture

Live video stream throughout, optional photo of ID at the start.

👤 Face detection

Detects no-face, multiple faces, and disabled camera — flagged with timestamps.

🖥️ Screen recording

Optional full-screen capture with candidate consent and S3-stored evidence.

🎙️ Voice detection

Picks up audio in the room — catches verbal cheating, not just visual.

🛠️ Devtools block

Detects browser devtools, console, and view-source — and shuts the exam.

↔️ Tab-switch & focus

Tracks tab changes, window switches, and copy-paste with severity scoring.

🖱️ Right-click & copy

Right-click, copy, and paste disabled per exam — gracefully, not aggressively.

⛶ Fullscreen lock

Mandatory fullscreen on start; exiting triggers a warning or auto-submit.

⚠️ Auto-submit threshold

Configurable per exam — N violations of severity X and the exam closes itself.

👁️ Live proctor warnings

Real-time messages to the candidate from a human proctor via Socket.io.

Every violation is timestamped, severity-classified (low / medium / high / critical), and saved to a post-exam review queue. A human decides — it's never a silent auto-fail over a flaky webcam.

Live monitoring dashboard

Watch the whole room. Without being in the room.

Every active candidate, on one screen. Live webcam thumbnails, real-time violation alerts, and a one-click warning the candidate sees instantly — for as long as the exam is running.

Live drive · Senior Frontend Engineer LIVE

89Active

78On track

9Warnings

2Critical