High-Level Overview#

Uniplaner is a modern academic scheduling and planning system designed for higher-education institutions. The product supports the full lifecycle of timetable creation - from collecting authoritative academic data and availability constraints, through optimization and conflict detection, to publishing finalized schedules back into university management systems.

The system enables universities to centrally manage academic calendars, time patterns, room inventories, instructor availability, and institutional constraints, and then automatically generate high-quality, conflict-free schedules. It supports multiple user roles (planners, administrators, teaching staff etc.) and allows both automated optimization and manual fine-tuning of schedules.

Uniplaner integrates directly with institutional systems such as USOS to both consume source data (courses, groups, instructors, rooms) and publish finalized timetables, ensuring consistency with the university’s core systems.

My Responsibilities#

This is the first major product fully led and implemented by me as a technical owner. I am responsible for the entire backend architecture and business logic, including system design, performance, and integration.

My responsibilities include:

• Design and implementation of all backend services using Kotlin and Ktor.
• Schedule generation and optimization logic, including constraint handling and prioritization.
• Conflict detection across rooms, instructors, student groups, and time slots.
• Real-time plan updates supporting multiple concurrent editors.
• Seamless, near-real-time integration with USOS for data synchronization and plan publishing.
• Report generation in multiple formats (PDF, CSV, XML, JSON, XLSX).
• Full-text search over imported USOS data to improve planner productivity.
• Query and data-access optimization, with most endpoints - including complex ones - responding in under ~700 ms.
• Requirements analysis and close collaboration with domain experts from universities.

System Characteristics#

From a technical perspective, Uniplaner is characterized by:

• Event-driven communication using Apache Kafka to propagate plan updates.
• Strong consistency and transactional guarantees backed by Oracle.
• Schema versioning and controlled database evolution using Liquibase.
• Type-safe SQL and data access implemented with JetBrains Exposed.
• Centralized authentication and authorization via CAS.
• Containerized deployment using Docker.
• A modern web frontend built with React and TypeScript, consuming backend APIs in real time.

Scheduling Workflow (Conceptual)#

The typical scheduling process supported by the system includes:

1. Import of courses, groups, instructors, and rooms from the university’s management system.
2. Collection of instructor availability and institutional constraints.
3. Definition of planner preferences and priorities (e.g., room types, block scheduling).
4. Automated generation of an optimized timetable.
5. Manual adjustments, validation, and conflict resolution where needed.
6. Publishing of the finalized schedule back to USOS or another authoritative system.

Key Outcomes#

• Significant reduction in manual effort required to build and maintain academic schedules.
• Improved schedule quality through automated optimization and early conflict detection.
• Real-time collaboration for planners and administrators.
• High system responsiveness despite complex scheduling logic.
• A reusable scheduling platform adaptable to different universities and academic rulesets.

This project demonstrates end-to-end product ownership, combining domain modeling, performance-oriented backend engineering, and complex integration with external institutional systems.