AI-Powered Apps

AI-powered apps / case 01

OptiTalent

Evidence-first recruiting workflows

A multi-tenant recruitment product that turns candidate search, evidence review, shortlist decisions, controlled sharing, and recruiter feedback into one coherent workflow.

Classification
Independent Project
Delivery status
Independent product / active validation

The problem

Recruiting teams need to move quickly without turning matching into an opaque score or losing the evidence behind a decision.

System arrangement

How the system works

Confirmed role constraints fan into exact and semantic recall lanes. Candidate-level RRF builds the pool, ColBERT refines it, and evidence policy controls what can reach a recruiter shortlist.

  1. 01
    SourceRole intent + constraints

    Confirmed eligibility compiles first

  2. 02
    CapabilityPGlite exact + lexical

    Canonical facts, filters, and FTS

  3. 03
    CapabilityQdrant dense + sparse

    BGE-M3 versioned derived index

  4. 04
    CapabilityCandidate-level RRF

    Fuses only executed recall lanes

  5. 05
    CapabilityColBERT MaxSim

    Reranks the fused candidate pool

  6. 06
    ControlEvidence policy

    Unknown stays unknown, not excluded

  7. 07
    OutcomeRecruiter shortlist

    Evidence remains reviewable

  8. 08
    ControlTenant-local compute

    Redacted subjects stay close to data

My role

I defined the product boundaries, tenant model, evidence contracts, recruiter workflow, review gates, and release acceptance. AI agents produced most implementation code under those constraints.

Verification evidence

  • Public product surface at app.optitalent.cc.

  • Implemented candidate search, candidate evidence, shortlist, and sharing contracts.

  • Tenant-scoped retrieval and explicit review states are present in the product runtime.

  • LightGBM LambdaRank was evaluated against deterministic baselines on a frozen retrieval protocol and was not selected as the production winner for that feature shape.

Safety and architecture decisions

  1. 01

    Prefer evidence-first retrieval and explicit unknowns over a single opaque LLM judgment.

  2. 02

    Place bounded embedding and retrieval work close to tenant data while preserving server-side policy and review authority.

  3. 03

    Keep the policy and lifecycle authority behind stable contracts while interfaces render approved state.

  4. 04

    Treat agent and human entry points as adapters over the same capabilities.

Capabilities

  1. 01

    Hybrid lexical and dense candidate retrieval

  2. 02

    Evidence-preserving reciprocal-rank fusion

  3. 03

    Local embedding and Qdrant recall runtime

  4. 04

    Candidate evidence and explainable match signals

  5. 05

    Draft and approved shortlist states

  6. 06

    Controlled sharing and feedback

  7. 07

    CLI, MCP, desktop, and web access paths