GistED
Your AI-powered learning companion—smarter studying, simplified
Created on 23rd March 2025
•
GistED
Your AI-powered learning companion—smarter studying, simplified
The problem GistED solves
GistEd is an AI-powered, all-in-one learning platform that simplifies education by centralizing study materials and enhancing productivity. It addresses the problem of information overload, helping students and lifelong learners stay organized and focused. With features like an Interactive AI chatbot, YouTube content extraction, community-driven library, task management, doodle board, handwritten note conversion, and NLP-powered chat for sorting and emailing documents. .
Main Highlights:
AI chatbot with memory and RAG for personalized learning.
Extract and summarize YouTube content effortlessly.
Community-driven library for sharing study resources.
Task management with calendar integration.
Interactive doodle board for brainstorming.
Convert handwritten notes to digital text.
NLP-based chat that intelligently sorts documents and emails them.
GistEd makes learning smarter, faster, and more collaborative and is your ultimate companion for a streamlined and efficient learning experience.
Challenges we ran into
- Email Automation (for Building the Lang Graph)
Challenge: Automating email workflows to send summarized content based on simple natural language queries was complex. Ensuring the system could interpret user queries accurately and deliver the right content without errors was a major hurdle.
Solution:
Built a natural language processing pipeline to interpret user queries and extract relevant content.
Implemented email automation to send summarized content directly to users.
Added error handling to manage edge cases like ambiguous queries or failed deliveries.
- Agentic AI (Building Custom Tools)
Challenge: Creating custom tools for fetching question papers and study materials required building a system that could understand user requests, search across diverse sources, and retrieve accurate results.
Solution:
Developed custom tools that integrate with multiple data sources to fetch relevant materials.
Designed a feedback mechanism to improve the accuracy of fetched results over time.
Ensured the system could handle complex queries and provide precise results.
- Hybrid RAG Implementation
Challenge: Implementing a hybrid Retrieval-Augmented Generation (RAG) system faced challenges with reranking results. Combining dense and sparse retrieval methods often led to irrelevant or low-quality results being prioritized.
Solution:
Improved the reranking algorithm to better prioritize relevant
Conducted extensive testing to ensure the system delivered accurate and coherent responses.
- Semantic and Vector Search
Challenge: Implementing semantic and vector search to retrieve relevant study materials required handling high-dimensional data and ensuring fast, accurate results.
Solution:
Optimized the search algorithm to handle high-dimensional vector data efficiently.
Implemented caching mechanisms to speed up search results for frequently queried topics.