Talent AI
Talent AI reads between the lines of every CV — mapping real competencies, inferring career trajectory, and surfacing the candidates most likely to succeed in your specific role. Not just the ones who used the right keywords.
What recruiters actually see in the product.
Ranked candidates, reasoning, and confidence signals — not a metrics tile.
Turn role requirements into a structured scoring baseline
Talent AI reads the job description, extracts real hiring criteria, and turns vague recruiter language into a scoreable role profile.
Two workflows. One intelligence layer.
Talent AI operates in two modes that share the same underlying intelligence — a recruiter-facing pipeline for evaluating and shortlisting candidates, and a candidate-facing workspace for coaching, self-assessment, and application preparation.
JD Analysis
Paste your job description. Talent AI extracts required skills, infers must-have competencies, flags ambiguous requirements, and structures them into a scoreable role profile.
jd_analysis intent. Outputs a structured role spec used as the scoring baseline for every candidate evaluated against this req.
Knockout Question Generation
Instantly generate role-specific screening questions — skills-probing, situational, and knockout criteria that filter for actual job fit, not keyword presence.
knockout_question_generation intent. Questions are tuned to the role level, function, and specific requirements extracted in Step 01.
Bulk Resume Ingestion
Upload hundreds of resumes at once — PDF, DOCX, plain text, or LinkedIn exports. Talent AI deduplicates, parses, and queues every candidate for parallel screening.
Batch upload via dashboard or S3/GCS sync. Each resume is parsed into a structured profile with 200+ extracted attributes.
Parallel Prescreening
Every candidate is scored simultaneously across four dimensions: overall fit, skill match, experience quality, and role alignment — each with a 0–100 score.
candidate_prescreen intent. Produces overall_score, fit_score, skill_score, experience_score plus recommendation: strong_yes | yes | maybe | no | strong_no.
Ranked Shortlist
Candidates surface ranked by recommendation tier. Every decision is explainable — recruiters see the exact signals, strengths, concerns, and risk flags that drove each score.
Bias guardrails run on every scoring event. Shortlist output includes per-candidate rationale, strengths[], concerns[], and risk_flags[] arrays.
Interview Question Generation
For each shortlisted candidate, generate a tailored interview question set — tuned to their specific skill gaps, notable strengths, and the exact role requirements.
interview_question_generation intent. Questions are personalized to the candidate's profile, not generic role templates.
Human Review & Override
Recruiters approve, adjust, or override AI recommendations. Every decision — AI-generated or human-adjusted — is logged in a complete, GDPR-compliant audit trail.
Fields: human_reviewed, human_override, reviewer_notes. Full audit log per candidate per role. Candidate scoring using candidate_scoring intent for manual adjustments.
Resume Analysis & Improvement
Structured breakdown of resume strengths and weaknesses — with specific rewrite suggestions and positioning improvements for your target role.
Job Fit Analysis
Paste any job description to see a skill-by-skill match against your background. Understand exactly where you meet the bar and where you fall short.
Skill Gap & Career Assessment
See where you stand today vs. your target role. Get targeted, time-boxed learning paths to close each gap with a clear sequence.
Career Path & Transition Planning
Map realistic trajectories — lateral moves, industry transitions, and promotion paths — with step-by-step action plans grounded in your actual profile.
Mock Interviews & Coaching
Practice role-specific interview scenarios with real-time feedback. Coaching sessions remember your full context and progress across every conversation.
Application Materials
Generate tailored cover letters and recruiter intro scripts — tuned to the specific role, company context, and your unique background.
Deploy the way your team works
From a no-code dashboard to fully self-hosted infrastructure — Talent AI fits into your existing stack, not the other way around.
One call. Full intelligence.
Score a candidate against a role with a single API call. Guardrails run automatically — bias detection, PII redaction, audit logging.
curl -X POST https://api.liyaengine.com/v1/run \
-H "x-api-key: $LIYA_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"pack": "talent-intelligence",
"intent": "score-candidate",
"input": {
"resume_text": "Sarah Chen — 6 years building data pipelines...",
"role": {
"title": "Senior Data Engineer",
"requirements": ["Python", "Spark", "dbt", "SQL", "cloud infra"],
"level": "senior"
}
},
"guardrails": {
"bias": { "policy": "eeoc-compliant", "action": "flag" },
"pii": { "action": "redact", "fields": ["name", "email", "address"] }
}
}'{
"response": {
"candidate_id": "cand_8f3a2b",
"fit_score": 0.87,
"skill_match": {
"matched": ["Python", "Spark", "SQL", "GCP"],
"inferred": ["dbt", "pipeline orchestration"],
"gap": ["Kafka"]
},
"trajectory": "strong upward — 3 promotions in 6 years",
"coaching_insight": "Kafka gap closeable in ~6 weeks; strong RAG signal",
"guardrails": {
"bias_flags": [],
"pii_redacted": ["name", "email"],
"audit_id": "aud_7c2d9f"
}
},
"execution": {
"steps": 4,
"latency_ms": 1240,
"agent_mode": "agentic"
}
}Liya Engine components powering Talent AI
Ready to find better candidates?
Talent AI is live. See the scoring engine in action on your own job reqs.