This innovative platform offers an immersive way to explore and manage NFTs, with features like wallet reports, collection tracking, and creator insights. Built with React.js, Node.js, and GraphQL, it leverages AWS for scalable cloud infrastructure and integrates a robust blockchain index using Cassandra, MongoDB, and Ethereum to efficiently handle over 20 million transactions.
This project involved building a specialized TikTok scraper to extract fashion-related content, utilizing Python and Selenium for web scraping and MongoDB for data storage. The project assigns relevance scores to posts using machine learning and NLP, ensuring that only the most relevant fashion content is retained. Deployed on AWS with a scalable MongoDB Atlas cluster, this solution balances efficiency with cost-effectiveness.
This full stack chat application, built using React, Spring Boot, and WebSockets, features a robust architecture with microservices orchestrated through Eureka and Docker Compose. The project includes extensive logging, error handling, and both unit and integration testing for frontend and backend services, ensuring reliability and scalability in a containerized environment.
This proprietary task management application, inspired by Jira, is built using React, Node.js, and MongoDB. It provides a streamlined interface for managing tasks, tracking progress, and enhancing team collaboration. The project focuses on delivering a robust and intuitive tool tailored to the needs of modern development teams.
This project involves developing a multithreaded Elastic Search server in Go, designed to enhance search query performance on MongoDB. By leveraging Go's concurrency features, the server delivers efficient and fast search capabilities through its robust API, making it ideal for high-demand applications.
Proposed a probabilistic constant time solution to reduce high read latency for high-read/low-write data in LSM Tree based Rocks DB, an embedded key-value store. We solve this problem using probabilistic data structures(Bloom filters).
This project involves designing a robust blockchain indexer that leverages Cassandra, MongoDB, Hadoop, Ethereum, and Redis to efficiently store and query over 20 million blockchain transactions, including more than 1 billion NFT and ERC20 transfers. The solution features distributed data pipelines built with Airflow, Spark, and Node.js, capable of processing millions of daily operations.
This project involves building a sophisticated search engine using Java and the Lucene library, with a focus on object-oriented design and custom ranking algorithms. The search engine parses and indexes documents, implements multiple ranking strategies including TF-IDF and relevance feedback, and introduces a novel optimized TF-IDF algorithm that enhances search accuracy by weighting query terms based on their relevance.