Data / 2024
Campus FAQ Chatbot
Less searching. More answers. A focused language assistant for campus questions.
The problem
Recurring campus questions are answered across scattered sources. A small, focused assistant makes those answers easier to retrieve.
My role
NLP pipeline, backend API, evaluation.
An NLP-based FAQ assistant using Flask, NLTK, and TF-IDF similarity matching.
The approach
- Organized campus FAQs into a structured knowledge base and preprocessed queries with NLTK.
- Used TF-IDF and cosine similarity to match questions to relevant answers.
- Exposed the workflow through Flask and evaluated it on curated campus prompts.
The toolkit
The outcome
The checked-in resume reports 85%+ accuracy on test queries from a curated campus FAQ dataset.
The reported accuracy is project-specific, not a general language-understanding benchmark. No production usage or independent evaluation is claimed.
Keep exploring