About UET GPT
The mission, architecture, and student community behind the AI assistant for UET Taxila.
Last updated: 2026-09-25
Our Mission
UET GPT was created to solve a persistent challenge faced by thousands of prospective and current students at the University of Engineering and Technology (UET), Taxila: navigating complex admissions policies, fee structures, academic regulations, and campus announcements scattered across disparate documents and portals.
How It Works
UET GPT employs Retrieval-Augmented Generation (RAG). When a user submits a query, our retrieval system searches official UET Taxila documentation (including undergraduate prospectuses, department bulletins, and fee schedules) and supplies this verified context to our language models. This prevents hallucinations and ensures all answers are backed by verifiable citations.
Community & Independence
UET GPT is an independent, open-source initiative built by students and alumni. It is not officially operated or endorsed by UET Taxila. We believe in transparent, accessible knowledge for every engineer and student.
How We Verify Facts
Answers are retrieved from a knowledge base built out of official UET Taxila documents - prospectuses, fee schedules, department pages, admission notices and hand-checked reference files - and every answer links back to the documents it used. When a figure comes from a document rather than a live university source, we label it as indicative instead of presenting it as an official announcement.
Each page carries a "Last updated" date derived from when its content last changed. If you spot something that is out of date or wrong, report it on the public issue tracker - it is the same channel the repository uses for corrections, and the commit history is the public record of what changed and when.
Founder
UET GPT is founded and maintained by Ibrahim Salman, who built the RAG pipeline, crawler infrastructure, and calculator tools that power the platform. The project's source code is public on GitHub under the AGPL-3.0 license, so its retrieval logic and data sourcing can be independently verified. See more of his work at ibrahimsalman.vercel.app.