Tapestree
Cloud-based institution management for Nigerian universities. AI-generated timetables, JAMB integration, and the kind of data security that meets NDPR compliance.
- Practice
- BUILD
- Year
- 2024
- Stack
- Next.jsPythonPostgreSQLOpenAIAWS
- Live
- tapestree.co
At a glance
to generate a clash-free 5-day timetable for 800 students
previous manual process, now compressed
AES-256 encryption, audit logging, role-based access
Universities were spending 200+ hours per semester on timetables — by hand.
Nigerian universities run on a different operational rhythm than universities in most other countries. Course catalogs are larger. Lecturer-to-student ratios are tighter. Venues are scarcer. And the timetable build process — assigning lecturers, courses, students, and rooms across a 5-day week — was happening in spreadsheets, with manual reconciliation, for weeks before each semester.
Tapestree came to us with a vision: AI-driven timetable generation that would handle the constraint problem in minutes, not weeks. But it needed to do more than that. It needed to be a complete institution management system — payments, JAMB integration for admissions, NDPR-compliant data handling, and the rest.
The AI isn't the model. It's the constraint mapping.
The common mistake with AI features is reaching for a large language model when what you actually need is a constraint solver. We started by mapping the timetable problem properly — hard constraints (a lecturer can't be in two rooms simultaneously), soft constraints (a lecturer prefers afternoons), and optimization targets (minimize gaps for students).
The first iteration used a constraint-satisfaction algorithm augmented with heuristic search. Later iterations layered in LLM-powered features for natural-language schedule queries — “What rooms are free Thursday at 2pm?” — but the core scheduling engine doesn't depend on an LLM.
Parallel to the scheduling work, we built out the institution management spine: student records, course catalog, lecturer profiles, attendance tracking, payment processing through three Nigerian gateways, and NDPR-compliant audit logging.
What we built.
- AI Timetable Engine — Constraint-satisfaction scheduler with optimization heuristics. Clash-free 5-day timetables for departments of 800 students generated in 12 minutes. Regeneration on rule changes happens in seconds.
- Student & Faculty Records — Complete academic records management. Student lifecycle from admission through graduation. Faculty profiles with course loads and availability.
- JAMB Integration — Direct integration with the Joint Admissions and Matriculation Board for verifying student credentials during admissions.
- Multi-gateway payments — Paystack, Flutterwave, and Remita integrations with automatic gateway selection based on transaction amount and student preference.
- Security & Compliance — AES-256 encryption at rest. Role-based access controls. Audit logging for every sensitive operation. NDPR-compliant data handling.
Outcomes
- Timetable generation time reduced from 3 weeks to 12 minutes.
- Regeneration on disruption (new lecturer, venue change) — seconds instead of days.
- NDPR compliance achieved and audited.
- Three Nigerian payment gateways integrated with automatic routing.
- Continuous development relationship with feature additions monthly.
Tech stack
They turned a three-week scramble every semester into a twelve-minute task. The engineering held up under real load.