Responsible AI Decisions
Sweepster is designed around trust, reversibility and clear boundaries between known facts, inferred evidence and user intent.
No imputation
Unknowns are never guessed into the score. A fabricated score is worse than no score.
Human-in-the-loop
The system asks questions only when the answer can materially change the score or build the profile.
Reversible learning
Feedback rules are proposed, visible and reversible. Learning is not hidden personalisation.
Auditability
Requirements preserve source evidence. Hallucinated requirements are rejected before scoring.
Sweepster is not trying to automate a person’s career decision. It makes the decision more structured, explainable and less dependent on keyword luck.