Table of Contents

Code ↗

I built a public scheduling tool for Emory University’s Residence Life staff, reducing manual rota construction from hours to seconds.

The problem

Staff scheduling looks simple until you write down the rules: people have availability windows, every shift needs coverage, and the workload should be fair. Done by hand, it took hours and still left someone unhappy.

The approach

I modeled availability, coverage, and fairness as a constrained combinatorial optimization problem and solved it with warm-start stochastic local search — start from a reasonable guess, then make small improving swaps until the schedule satisfies the constraints and balances the load.

Outcome

• Cut schedule construction from hours to seconds.
• Shipped as a tool non-technical staff could actually use.

Add the repository link and a screenshot here.