Nutrition Research Chatbot

Project: Nutrition Science Chatbot

Situation

Users needed a faster way to search a large body of nutrition content and ask questions in natural language.

Task

Build an early AI research assistant that crawled nutrition content, indexed it, and answered questions through Telegram and WhatsApp.

Action

Built a Scrapy pipeline to collect nutrition articles
Used a fine-tuned Hugging Face question-answering model against the collected content
Connected the service to Telegram and WhatsApp through Twilio
Kept an offline sample dataset for repeatable development and testing

Result

Created an accessible question-answering interface for nutrition information and an initial content-ingestion workflow that could be extended into retrieval-augmented generation.

Technologies

Python, Scrapy, Hugging Face Transformers, Telegram, WhatsApp, Twilio