Liya Engine

AI-powered recruiting intelligence — screen, score, and shortlist candidates at scale using the full Hiring domain pack.

Talent AI is Liya Engine's solution for recruiting teams. It replaces manual resume review with structured AI intelligence that understands careers, competencies, and role fit — not just keyword matching.

Domain pack: hiring
API base path: POST /v1/hiring/{intent} or POST /v1/run with pack: "talent-ai"

What Talent AI does

Capability	What it replaces
Parse and score every resume against a JD	Manual resume review
Surface hidden competencies from career history	Keyword filter screens
Generate ranked shortlists with explainable rationale	Recruiter gut-feel ranking
Generate role-specific knockout and interview questions	Generic question banks
Coach candidates on gaps, career direction, and interview prep	Ad-hoc feedback

Recruiter workflow

The typical recruiter pipeline runs five intents in sequence:

Step 1 — Analyse the job description

POST /v1/hiring/jd-analysis

{
  "input": {
    "user_id": "recruiter_001",
    "job_description": {
      "title": "Staff Engineer",
      "description": "We are looking for a staff-level engineer to lead our platform team...",
      "company_name": "Acme Corp",
      "required_experience_years": 7
    }
  }
}

Returns a structured role spec — required skills, inferred competencies, seniority signals — used as the scoring baseline for every candidate.

Step 2 — Generate knockout screening questions

POST /v1/hiring/knockout-question-generation

{
  "input": {
    "user_id": "recruiter_001",
    "job_description": {
      "title": "Staff Engineer",
      "description": "..."
    }
  }
}

Returns a set of role-specific knockout questions tuned to the seniority level and technical requirements of the role. Use these in your application form or ATS to filter before bulk screening.

Step 3 — Prescreen candidates

POST /v1/hiring/candidate-prescreen

{
  "input": {
    "user_id": "recruiter_001",
    "profile": {
      "name": "Jane Smith",
      "title": "Senior Engineer",
      "years_of_experience": 6,
      "skills": ["TypeScript", "Kubernetes", "Go"],
      "work_experiences": [
        {
          "title": "Senior Engineer",
          "company": "Stripe",
          "start_date": "2020-01-01",
          "end_date": null,
          "achievements": ["Led platform migration reducing deploy time by 60%"]
        }
      ]
    },
    "job_description": {
      "title": "Staff Engineer",
      "description": "...",
      "required_experience_years": 7
    }
  }
}

Returns a multi-dimensional prescreen result:

{
  "output": {
    "overall_score": 84,
    "fit_score": 88,
    "skill_score": 79,
    "experience_score": 81,
    "recommendation": "strong_yes",
    "strengths": ["Deep Kubernetes experience", "Track record of platform-scale impact"],
    "concerns": ["1 year short of required experience"],
    "risk_flags": [],
    "rationale": "Strong technical depth and measurable impact at a comparable scale..."
  }
}

recommendation is one of: strong_yes · yes · maybe · no · strong_no

Step 4 — Score and rank the shortlist

POST /v1/hiring/candidate-scoring

Runs a deeper comparative scoring pass on shortlisted candidates. Returns a ranked list with per-dimension breakdowns. Call this after prescreening to produce the final ranked shortlist your hiring team acts on.

Step 5 — Generate interview questions

POST /v1/hiring/interview-question-generation

{
  "input": {
    "user_id": "recruiter_001",
    "profile": { "name": "Jane Smith", "skills": ["TypeScript", "Kubernetes"], ... },
    "job_description": { "title": "Staff Engineer", ... }
  }
}

Returns a personalised interview question set tuned to the candidate's specific profile — not generic role templates. Questions target skill gaps, probe notable strengths, and cover the exact role requirements.

Candidate-facing workflow

Talent AI also powers candidate-facing features in career and job platforms:

Intent	Use case
`resume-analysis`	"Here's how strong your resume is and what to fix"
`resume-improvement`	"Here are specific edits to make your resume more compelling"
`cover-letter-generation`	"Generate a tailored cover letter for this role"
`intro-script-generation`	"Generate a 60-second intro for this interview"
`job-fit-analysis`	"How well does this role match your background?"
`career-assessment`	"Where are you in your career trajectory?"
`skill-gap-analysis`	"What skills do you need for your target role?"
`career-path-planning`	"What are your realistic next moves from here?"
`career-transition-planning`	"How do you move from X industry to Y?"
`mock-interview`	"Practice the exact questions you'll be asked"
`coaching-session`	"Ongoing career coaching conversation"
`general-chat`	"Free-form career Q&A"

Full intent reference

Intent	Endpoint	Required inputs
`resume-analysis`	`/v1/hiring/resume-analysis`	`resume`
`resume-improvement`	`/v1/hiring/resume-improvement`	`profile`
`cover-letter-generation`	`/v1/hiring/cover-letter-generation`	`profile`, `job_description`
`intro-script-generation`	`/v1/hiring/intro-script-generation`	`profile`, `job_description`
`jd-analysis`	`/v1/hiring/jd-analysis`	`job_description`
`knockout-question-generation`	`/v1/hiring/knockout-question-generation`	`job_description`
`interview-question-generation`	`/v1/hiring/interview-question-generation`	`profile`, `job_description`
`job-fit-analysis`	`/v1/hiring/job-fit-analysis`	`resume`, `job_description`
`candidate-prescreen`	`/v1/hiring/candidate-prescreen`	`profile`, `job_description`
`candidate-scoring`	`/v1/hiring/candidate-scoring`	`profile`, `job_description`
`career-assessment`	`/v1/hiring/career-assessment`	`profile`
`skill-gap-analysis`	`/v1/hiring/skill-gap-analysis`	`profile`
`career-path-planning`	`/v1/hiring/career-path-planning`	`profile`
`career-transition-planning`	`/v1/hiring/career-transition-planning`	`profile`, `career_goals`
`mock-interview`	`/v1/hiring/mock-interview`	`profile`
`coaching-session`	`/v1/hiring/coaching-session`	—
`general-chat`	`/v1/hiring/general-chat`	—

See Hiring Domain for full input/output schemas.

Knowledge sources

Upload your own data to ground Talent AI responses in your company's context:

Source type	Namespace	What to upload
Resume	`hiring:resume`	Candidate CVs for retrieval-augmented scoring
Job Description	`hiring:job_description`	JD library — past and active roles
Policy	`hiring:policy`	Hiring policies, EEOC rules, compensation bands
Profile	`hiring:profile`	Structured candidate profiles

# Upload a hiring policy document
curl -X POST https://api.liyaengine.ai/dashboard/knowledge/upload \
  -H "Authorization: Bearer $JWT_TOKEN" \
  -F "file=@hiring_policy.pdf" \
  -F "domain=hiring" \
  -F "source_type=hiring:policy"

Using the unified run endpoint

You can also call Talent AI via the unified /v1/run endpoint using pack:

POST /v1/run
{
  "pack": "talent-ai",
  "intent": "candidate-prescreen",
  "input": {
    "user_id": "recruiter_001",
    "profile": { ... },
    "job_description": { ... }
  }
}

This is equivalent to POST /v1/hiring/candidate-prescreen and returns the same response shape.

Next steps

All Hiring Intents — full input/output schema reference
Building a RAG Pipeline — upload your resume and JD library
Sessions and Memory — multi-turn coaching and interview flows
Configuration — persona, models, and guardrails

Talent AI

On this page