Assignment - all you need are abstracts.

Put articles' abstracts and reviewers' abstracts into .csv files. The algorithm will model the interests of reviewers and topics of articles and match them. You know additional conflicts of interests beyond co-authorship? No problem, upload an additional .csv file with a list of conflicts of interests.

Scoring - correct for systematic harshness.

Upload the list of reviewer scores and the algorithm will try to estimate the underlying article's score. Have overly hash or soft reviewers? The algorithm will appropriately correct those cases. A reviewer always disagrees with other reviewers and another always agrees with them? The algorithm will differently weigh their scores. Oh, and the algorithm will tell you how uncertain it is.