BlindMatch PAS
A fair, preference-driven Student–Supervisor matching system.
Matching that removes bias from the loop
BlindMatch PAS is a project allocation system that pairs final-year students with academic supervisors based on ranked preferences and capacity constraints, rather than first-come-first-served. Built with ASP.NET Core 9.0, it exposes an admin console for running allocation rounds and a student portal for submitting weighted preferences.
Manual allocation was slow and contested
The existing process was a spreadsheet negotiated over email — opaque, easy to game, and impossible to audit. Students with quieter voices lost out, and supervisors were routinely over- or under-allocated. We needed a deterministic, explainable algorithm with a clean paper trail.
Leading a 9-person team to a shippable core
As team lead I owned the architecture, sprint cadence, and the allocation engine. We modelled the problem as a constrained stable-matching variant, wrote it test-first, and layered a role-based portal on top. I ran weekly integration reviews to keep nine contributors converging on one coherent codebase.
Reproducible rounds, in seconds
Allocation that took days of back-and-forth now runs in seconds and produces an auditable result every stakeholder can inspect. The blind, preference-weighted model measurably improved first-choice match rates while keeping supervisor loads within capacity.