Talentlo Intelligence
The brain behind
the shortlist.
Mark 1.0 is a composite AI model for hiring intelligence, built for the JD-to-shortlist pipeline. It reads the role, reads the evidence, and returns a ranked Top 30 with the reasoning behind every line.
Built in India · DPDP aligned
Inside Mark 1.0
From job description to a ranked shortlist.
Five steps, one outcome. The recruiter decides; Mark does the reading, scoring, and explaining.
Role context
What does this job require?
Skills, seniority, domain, salary band, notice tolerance. Structured from the JD before the first application arrives.
Candidate evidence
What has this person demonstrated?
Each profile is read as an evidence trail: capability, career trajectory, tenure patterns, availability signals.
Signal scoring
How does evidence map to the role?
Candidates are evaluated across 7 dimensions. Weights adjust to role context and hiring constraints.
Decision intelligence
What does this mean for the decision?
A verdict, evidence citations, and the single biggest thing to verify before the interview. Per candidate.
Decision packet
Ranked shortlist, evidence-backed.
Top 30: composite score, 7-dimension breakdown, evidence trail, one risk flag per candidate.
7 scoring dimensions
Experience, Skills and Growth carry higher weight across most role types.
Model specification
Mark 1.0 at a glance.
The full spec: what it takes in, what it produces, and where the boundary of the decision sits.
talentlo/mark-1.0
Composite AI: multiple AI techniques working together for a single hiring outcome
Candidate evaluation · Hiring decision support
Role understanding · Candidate evaluation · Fit scoring · Decision output
Job descriptions · Candidate profiles · Hiring constraints · Pipeline context
Decision Packet: ranked candidates, evidence citations, 7-dimension scores, risk flags
7 dimensions: Experience, Skills, Salary Fit, Preference, Growth, Retention, Notice Period
Evidence-grounded: every score linked to specific candidate signals
Recruiter decision support · Human-in-the-loop
Native intelligence layer inside Talentlo · Not a standalone API
Built and deployed in India · Nothing leaves India · DPDP Act aligned
Mark 1.0 · Internal production · 2026
Evaluation
Measured on hiring, not on general AI benchmarks.
Mark is evaluated on metrics that map to hiring work. Results are published as data matures, not before.
Shortlist Precision
Percentage of Mark's Top 10 candidates that advance to a first-round interview. Measures signal quality, not volume.
Recruiter Agreement Rate
Correlation between Mark's ranked recommendations and recruiter selections. Validates hiring-domain accuracy.
Evidence Coverage
Percentage of score components directly mapped to verifiable candidate signals. Measures explainability depth.
Time-to-Shortlist
Hours saved from first application to a ranked Top 30, compared to manual screening.
Benchmark methodology: Talentlo HireEval v1 · Results published as evaluation data matures.
Model lineage
The Mark series.
Mark 1.0 is the first production generation. Later versions are announced when they ship. Never before.
Composite AI hiring intelligence: role-context engine, 7-dimension candidate scoring, evidence-grounded evaluation, and Decision Packet output. Powers every role posted on Talentlo.
Next versions are announced when they ship.
About the model
Technical questions, answered plainly.
Can’t find what you need? Talk to the team.
Contact the teamWhat does "Composite AI" mean for Mark 1.0?
Composite AI means Mark 1.0 combines multiple AI techniques rather than relying on a single method. Each technique handles the part of the hiring problem it is best suited for: understanding roles, evaluating candidate evidence, scoring fit, and producing explainable output. The result is more structured and auditable than a single model doing everything.
Does Mark 1.0 make hiring decisions?
No. Mark 1.0 is a decision-support system. Every recommendation is evidence-linked and reviewed by a recruiter before any hiring action. The recruiter decides. Mark informs. This is not a legal disclaimer. It is the design principle.
How does evidence-grounded scoring work?
Each of Mark's 7 dimensions is evaluated against specific signals in the candidate profile, not inferred from job title or keyword presence. Experience is matched against role and seniority requirements. Growth is measured from career trajectory. Retention is assessed from tenure patterns. Every score has a source. A recruiter can ask why a candidate ranked where they did and get a specific, verifiable answer.
Why 7 dimensions and not a single score?
A single composite score collapses the reason for a recommendation into a number. Seven dimensions let a recruiter understand which specific aspect carries the risk. A candidate with a 92 composite but a 10 on Notice Period tells a very different story than a 92 across all dimensions. The breakdown is the value.
Why is this not a public API?
Mark 1.0 runs as an internal production model inside Talentlo. It is not offered as a standalone API today. The intelligence is delivered through the Talentlo product: job posting, applicant ranking, candidate search, and the Decision Packet.