Skip to content
NewCoding rounds written from each candidate’s CV

Screen every CV in minutes.
Shortlist on evidence.

TalentPulse AI reads every CV, scores it against your role’s must-have skills and weights, and shows exactly why each candidate ranks where they do. Then AURA-9 interviews your shortlist and hands you an evidence-linked dossier.

  • Every score explained
  • Candidate consent built in
  • Full audit trail
  • Bulk CV parsing
  • Explainable ATS scores
  • Adaptive AI interviews
  • Photo ID check
  • 15 integrity signals
  • Coding round in 7 languages
  • Evidence-linked dossiers
  • Candidate consent
  • Full audit trail
  • Hiring-partner network
  • Offers signed online
  • Fairness analytics
0weighted criteria behind every CV score
0%of scores explained, skill by skill
0integrity signals checked silently
0languages in the coding round
How it works

From a job description to a ranked shortlist in three steps

No spreadsheets and no keyword guesswork. Recruiters stay in control of every weight and every decision.

  1. Job description → role profile

    Define the role

    Paste or upload a job description. Title, experience, must-have skills and scoring weights are extracted for you to adjust.

  2. CVs → ranked shortlist

    Upload CVs in bulk

    Drop a batch of PDF, DOCX or TXT files. Every CV is parsed, matched against the must-haves and ranked by its weighted score.

  3. Interview → dossier

    Shortlist and interview

    Candidates above your bar are pre-selected. Invite them to a structured AI interview and receive an evidence-linked dossier.

Explainable screening

A score your hiring managers can trust

Each CV is scored on five weighted criteria. The breakdown, the matched skills and the gaps sit right next to the number, so a ranking is never a mystery.

  • Whole-term skill matching“Java” never matches “JavaScript”; aliases such as k8s → Kubernetes are understood.
  • Weights you controlSkills, experience, domain fit, education and keywords, tuned per role.
  • Missing skills called outEvery row shows what matched and what is missing, so nobody is rejected without a reason.
  • A shortlist bar, not a black boxSet the threshold per role; candidates above it are pre-selected for review.
Screen your own CVs
SKSara KimFrontend Engineer · 5 yrs · Remote88
Skills40%92
Experience25%86
Domain fit15%84
Education10%80
Keywords10%90
React Design systems Accessibility TypeScriptSSR
Strong match. Four of five must-haves present; ask about server-side rendering in the interview.
The platform

Everything after the CV lands, in one workspace

Screening is the start. TalentPulse carries each candidate through interviews, dossiers and hand-offs to hiring partners.

Explainable ATS scoring

A 0–100 score with the weighted breakdown, the matched skills and the gaps behind it.

Bulk CV parsing

PDF, DOCX and TXT, one CV or a whole batch, read in seconds.

Job description import

Turn a JD into a structured role profile with weights.

Adaptive AI interviews

Questions planned from the role and the CV, with follow-ups on what was said.

Photo ID and integrity

A photo ID check before the first question, then 15 silent integrity checks.

Hiring-partner network

Share consented shortlists with client companies and track every decision they make.

Coding round from the CV

A task written from the candidate’s skills, run against hidden tests in 7 languages.

Evidence-linked dossiers

Transcript, rubric scores and audit trail in one PDF.

Roles and permissions

Tick exactly what each recruiter and partner role may do.

Approvals and offers

Scorecards, approvals and offers the candidate signs online.

AI interviews for your shortlist

Structured first-round interviews, scored the same way every time

Shortlisted candidates take an adaptive interview with AURA-9 from any quiet room. Questions are planned from the role and the CV, follow-ups respond to what the candidate actually said, and every score links back to the transcript.

  • Five rubric scores plus integrity
  • Photo ID check before the first question
  • Candidates never see detection results
  • PDF dossier with a tamper-evident fingerprint
Invited to an interview? Sign in here
AURA-9 · AI interviewerFrontend Engineer · question 3 of 10Live
Walk me through the architecture of the most complex product you have shipped.
A Next.js monorepo with about 40 routes. We moved data fetching into server components and cut the client bundle by a third.
You mentioned server components. What broke first, and how did you find it?
Answer scored8.4 / 10Specific, owned the outcome, measured the result
Security and compliance

Built for teams that answer for every hiring decision

Consent, audit, retention and access control are part of the product, not an afterthought.

Candidate consentProfiles reach a company only after the candidate agrees.
Complete audit logWho did what, when and from where, exportable.
Retention and erasureAutomatic retention periods and on-request erasure.
Access controlHashed credentials, two-step sign-in, lockout and per-role permissions.

Candidate video is analysed on the candidate’s device. Scores support human decisions; they never make them.

Ready when you are

See your next shortlist ranked in minutes

Sign in to set up your roles and screen your own CVs.