Inspiration
Navigating immigration rules can be daunting due to their complexity and vast documentation. I was inspired to simplify this process by building a tool that organizes information and provides precise, tailored answers.
What it does
The Immigration Rules Assistant allows users to upload documents, categorizes them into topics, and enables users to select a category to get answers to their immigration-related questions.
How we built it
Database Creation: Built a Snowflake database to store immigration documents and categorized them. Document Chunking: Created a function to break large documents into smaller, meaningful chunks. Vector Database: Stored categorized document chunks in a vector database for efficient search. Streamlit App: Developed a user interface where users can select categories, input questions, and get precise answers.
Challenges we ran into
Chunking large documents without losing context. Ensuring efficient vector database integration for fast and accurate search. Creating a user-friendly interface for complex backend processes. Managing compatibility issues with Snowflake integration and Python versions.
Accomplishments that we're proud of
Successfully organized and categorized immigration documents for precise retrieval. Implemented seamless Snowflake integration with a powerful search feature. Delivered a user-friendly app within the hackathon timeline.
What we learned
How to efficiently process and store large documents in vector databases. Improved understanding of Snowflake’s features and Streamlit’s capabilities. Collaboration and problem-solving under strict time constraints.
What's next for IMMIGRATION_RULES_ASSISTANT
Expand document categories to include more immigration topics. Integrate advanced AI models for more contextual and conversational responses. Add real-time updates to handle the latest rule changes. Introduce multilingual support and build a mobile-friendly version.
Built With
- cortext-search-api
- database
- python
- snowflake
- streamlit
Log in or sign up for Devpost to join the conversation.