Selected work / KU-RISE Scheduling
Academic systems · Constraints · Validation Synthetic demonstration
Making a timetable work
Turning academic constraints into a testable scheduling component.
Many constraints. One coherent week.
A timetable must coordinate courses, cohorts, faculty, rooms and campuses at the same time.
The problem
A schedule can appear complete while containing clashes, unsuitable rooms or impossible travel between consecutive classes. KU-RISE narrows a broader ERP ambition to this specific problem. It combines an academic data model, spreadsheet ingestion, preflight diagnostics, constraint-based scheduling and independent validation. Separating time-slot allocation from room assignment reduces the size of the scheduling problem.
My contribution
The institutional problem framing, requirements and technical component developed with AI assistance — together with an explicit account of its demonstration limits.
Deliverables
- Python scheduling engine
- Academic data model and Excel loader
- Preflight diagnostics
- Independent timetable validation
- Demonstration workbook and integration roadmap
Process
Decisions that shaped it
Scope a usable component.
Focus on the scheduling engine and academic model within a wider integration plan. Keep its boundaries and input requirements explicit.
Diagnose inputs before solving.
Check missing data and capacity constraints before attempting a schedule. Explain why an input set cannot produce a usable result.
Validate independently.
Check the completed timetable separately from the solver. Test conflicts and practical constraints rather than assuming a generated result is correct.
Outcome
The supplied project documentation records a synthetic demonstration placing 381 weekly sessions with zero conflicts. This is a component-level demonstration result using KU-shaped sample data.
Takeaway
The ability to reduce a large transformation brief to a concrete problem, explain trade-offs, and distinguish demonstrated performance from operational impact.
Honest limits
The demonstration is not evidence of institutional adoption, measured time savings or a complete ERP. Real operational data, integration and a controlled institutional pilot remain future work.
What would come next
- Validate the academic model with institutional users
- Integrate authorised source data
- Evaluate schedules in a controlled pilot