Peer evaluation that is anonymous, statistically validated, and built to scale.
The Problem
Peer evaluation at scale breaks down without structure. Students grade inconsistently, bias creeps in when identities are visible, and there is no reliable way to flag or review evaluations that are clearly unfair. The result is that grades from peer assessment are rarely trusted — by students, instructors, or institutions.
What We Built
PES assigns unique anonymous IDs to every submission-evaluator pair, runs statistical anomaly detection across evaluation patterns, and routes flagged cases through a structured escalation workflow — from automated detection to TA review to teacher intervention. Final scores are aggregated using statistical methods that account for evaluator bias, making peer assessment both scalable and auditable.