Projects
AI-Powered NGO Claim Verifier
Technologies: Python, API, ML, FastAPI, NLP
Key Features:
- AI-Powered Verification System: Built a production-grade NGO Claim Verification API that performs automated authenticity checks using multi-source cross-verification (RBI IFSC DB and news sources).
- High-Performance NLP Engine: Developed a custom, zero-dependency NLP engine for entity extraction, disaster classification, location parsing, and financial detection, optimized for Indian NGO campaigns with sub-50ms extraction time.
- Parallel Verification Pipeline: Designed a parallel verification pipeline in FastAPI that validates bank details and correlates disaster-related information to generate a trust score.
- Production-Ready Architecture: Implemented a clean, scalable architecture with modular verification engines, trust scoring logic, structured logging, and serverless-optimized performance achieving under 2 seconds end-to-end.
Crisp
Technologies: ReactJS, TypeScript, Redux, Tailwind CSS, Git, NodeJS, API, OpenAI GPT-4
Key Features:
- Developed an AI-based web application allowing users to get summary of lengthy articles within seconds.
- Integrated advanced functionalities including Rapid API for AI summarization - OpenAI's GPT Model 4, browser history storage, Copy-to-clipboard functionality, RTK Query for API requests, & UI/UX interface.
Forest Fire Detection
Technologies: Python, API, ML, React, Computer Vision
Key Features:
- Developed and implemented ML model to efficiently detect and contain wildfires by processing 4000+ live images and extracting frames from previous fire videos, leveraging dynamic camera database for real-time updates, and highlighted data impact, resulting in 88% validation accuracy.
- Created a user-friendly front-end interface using React, enabling immediate fire detection notifications.
Sorting Algorithm Visualizer
Technologies: Java, Swing, IntelliJ IDEA
Key Features:
- Built app using Java and Swing, algorithmic problem-solving implementing Bubble, Selection, Insertion & Quick sort with enhanced UI for boosting algorithm selection improving real-time visualization by 99%.
- Features: Up to 200% speed acceleration selection, array regeneration, optimized multi-threading