🇧🇩 Built for Bangladesh's HR teams

Every candidate,
scored on your rubric.

One link: CV screening, then a live AI interview. Every score recomputed in code.

Request access See how it works
One link, CV to interview Scored in code Fair to every accent
topcandidate.pro/app
TopCandidate dashboard — ranked candidates, pipeline, and the scoring funnel
Dashboard preview
How it works

One link, both halves of the funnel

One standard per job — first on paper, then in conversation.

A CV screening
1
Post the job, share one link
Set your weights, share the apply link — or upload a batch.
2
Every CV scored on your rubric
Two passes, blended, recomputed in code.
3
Ranked in your dashboard
Full report per candidate, emails from your own address.
B The live AI interview

Then the interview
comes to them.

Clear the CV threshold, and the interview starts — same link, same sitting, in the browser.

A real conversation, not a form
Adaptive follow-ups, from your job's questions — identical for everyone.
Every question shown as text, always
Spoken and on screen — read as well as listen.
Typed answers where writing reveals more
The interviewer waits silently while they write.
Scored fairly, monitored honestly
Substance over fluency. Signals for review, never verdicts.
What it does

The whole funnel, honestly built

Seven things it actually does. We run our own hiring on every one.

Fair to every candidate
Substance scores. English polish doesn't. A sharp answer in broken English beats a fluent empty one.
Tested on every scoring change
One link, full funnel
Apply, CV, interview — one link, one sitting.
AI CV screening
Six weighted dimensions. Scored in code, not guessed.
Live AI interview
Adaptive spoken + typed questions, drawn from your job.
Honest proctoring
Monitored and flagged for your review. Never called cheat-proof.
Full audit trail
Every score shows its evidence.
Built for Bangladesh
PDF batches, Gmail sends, works on a phone.
What actually runs

One CV, six checks, two opinions

Parsed separately, scored twice, checked by deterministic code.

In parallel
Parse CV
roles, skills, dates, education
Parse JD
requirements, must-haves
Two-pass scoring
Strict pass
gatekeeper · temp 0.2
Upside pass
hiring manager · temp 0.3
Blend 60 / 40 per dimension
Deterministic — no model call
Hard-gap check
−15 to −25 each, capped −35
Confidence gate
>12 pts apart → downgrade
Final weighted score
recomputed in code, not taken from the model · 0–100 + recommendation

Every stage after the blend is ordinary code, not a prompt — the same input always produces the same adjustment.

The rubric

Six dimensions, weighted for the role

Each scored 0–20. The weights are yours to change.

Skills Match25%
Experience Relevance20%
Achievement & Impact20%
Role Alignment15%
Education & Certifications10%
Stability & Tenure10%

Eight presets ship with the app — or set all six yourself.

Real internal numbers · August 2026

We built it for our own hiring first

Every application at LinkX360 and Data Solution 360 — the two companies behind it — runs through this exact funnel.

292
CVs screened
on the rubric
12
roles screened
across both companies
29
applications through
the public apply link
2
companies —
both of them ours

Honest small numbers beat invented big ones.

Access

Early access, by conversation

There is no checkout and no plan to pick. We're onboarding teams one at a time while the product is still early.

Contact us for access

Tell us what you're hiring for. If it's a fit, we set you up and walk the first batch with you.

Your own account, set up with you on a call
Weightings and interview questions tuned to the roles you actually hire for
Candidate emails sent from your own address
Direct line to us while you're getting started
Request access

Already set up? Sign in here.

Questions

Straight answers

What is the live AI interview?
A real-time spoken conversation in the browser — your job's questions, identical for every candidate, every question also shown as text. Some are answered in writing.
Is scoring fair to candidates whose English isn't fluent?
The scorers grade substance, not fluency — a sharp answer in broken English must outscore a fluent empty one, verified with test transcripts before every scoring change.
Does the proctoring catch cheating?
It produces signals for human review — monitoring, flagged snapshots, a coverage timeline. Nothing is auto-rejected, and no browser system is cheat-proof.
Does a missing requirement reject a candidate?
No. It applies a capped penalty and pushes the candidate down the ranking — never off it.
What happens when the two scoring passes disagree?
More than 12 points apart: flagged low-confidence and downgraded to Maybe. Never upgraded.
Where does candidate data go?
CV text and transcripts go to OpenAI's API for scoring — we say so plainly. PDFs delete after 30 days, and you can delete any screening.