A modern academic information system that streamlines administration, course enrollment, learning workflows, and verified AI-powered academic guidance.
SIAKAD AI was designed for medium-to-large higher education institutions in Indonesia. The platform supports students, lecturers, and faculty-level administrators across study-plan enrollment, grading, learning materials, attendance, thesis tracking, and academic records.
Legacy academic portals often lack responsive workflows, real-time validation, and reliable student support. Academic advisors are frequently overloaded with repetitive questions about credit limits, prerequisites, graduation readiness, and thesis procedures.
We built a complete academic information system with a grounded AI Academic Advisor. Instead of relying on generic model memory, the assistant receives a verified database snapshot of the student's academic profile before answering.
The system uses a dual-guard architecture: a grounded context builder compiles live student, curriculum, schedule, and grade data; then a validation engine checks the AI response for invalid course references, credit math errors, out-of-scope topics, and speculative language before the answer reaches the student.
Designed the database schema, implemented the LLM integration layer, built the context builder and validation pipeline, developed KRS credit calculations, enforced faculty-scoped security boundaries, and authored feature tests for core academic and AI workflows.
The guard engine prevents invalid course-code references and credit calculation errors, reducing misleading academic advice. Automated KRS validation also reduces invalid submissions and helps advisors process approvals faster.
We can help design academic workflows, internal portals, and AI assistants that stay grounded in real institutional rules.
Discuss Project