This page summarizes the baseline controls CandiLyt applies when processing candidate data.
CandiLyt is designed for decision support. Hiring teams must review shortlists, interview materials, feedback, and negotiation guidance before taking action. The product must not automatically reject, select, or advance a candidate without a human decision.
Workspaces are designed around organization isolation, fixed user roles, job and interview assignments, private notes, and default deny access. Sensitive hiring notes and negotiation guidance should be limited to people who need them.
CVs and candidate records should be stored privately, not exposed through public links. Access to restricted records should be checked on the server and logged where appropriate.
CandiLyt's initial policy is not to use customer candidate data, interview feedback, or hiring outcomes to train shared models.
CandiLyt should not infer protected characteristics, biometric identifiers, health or disability status, caste, religion, sexuality, pregnancy, emotion, personality, or similar sensitive traits for hiring decisions.
Hiring teams should see why a candidate appears relevant, what looks unclear, what must be checked in an interview, and who made the final decision. Scores should not be treated as standalone hiring decisions.
CandiLyt may use carefully reviewed third-party providers to operate and secure the service. The public Service Providers page explains the approach without exposing implementation details.
Security concerns can be reported to security@candilyt.com. CandiLyt will maintain an incident response process covering triage, evidence preservation, customer notification, and regulator reporting where required.