The educator describe their homework assignment in great detail, describing for the prospective grader everything they would need to know about the assignment, including guidelines for feedback and scoring, rubric or at least criteria for scoring, dates when papers will reach (and must return from) Gradify, policies for partial credit and late submissions etc.
Gradify's AI-based matching algorithm digests a wide array of Grader attributes, including education, work experience, the academic discipline of the grader's background, portfolio of prior grading work, grader rating by instructors, and availability to complete the job at hand. The algorithm gets more accurate over time, as it learns from past matches.
The grader, trained to follow sound pedagogical practices, embeds student assignments with detailed, constructive and actionable feedback. Assignments are scored on an instructor-approved rubric. If the educator is not satisfied, for whatever reason, Gradify will assign a new grader or refund the fee.