🚀 Passionate AI Engineer and Fullstack Developer who loves building impactful products.
🎯 Focused on AI/ML, scalable backend systems, and health-tech + blockchain innovations.

• 🔭 Currently working on: AI-powered apps, federated learning, IoT-ML systems
• 🌱 Exploring: MLOps, Cloud-native architectures, Advanced NLP
• 🤝 Open to collaborate on: AI/ML research, fullstack apps, open-source projects
• 🧠 Always curious: Hackathons, system design, and creative problem-solving

A simulation-based project that models the Federated Learning Mechanism using mathematical algorithms with blockchain and encyption.  

• Problem: Training a single, powerful model from data distributed across many different sources without ever seeing or centralizing the raw data.

• Solution: A federated learning system that orchestrates secure, collaborative training.

• Architecture Diagram: A central Flower server distributes a global model to multiple clients. Each client trains on its local data partition. The clients then send updated parameters (gradients) back to the server, which aggregates them and sends a new global model back to the clients. This process repeats for multiple rounds.

• Tech Stack: Python(FLOWER), NumPy, Matplotlib

• Highlights: Research-driven, visual simulations for biological processes.

A Chrome extension that provides AI-powered hints, explanations, and optimized solutions while solving problems on LeetCode.  

• Problem: The need for a simple, non-disruptive way to get AI assistance for LeetCode problems without switching tabs or complex interfaces.

• Solution: A minimalist Chrome extension that seamlessly integrates into the LeetCode environment.

• Visual Presentation: A floating AI icon appears in the bottom-right corner of any LeetCode problem page. When clicked, it expands to a menu of AI features such as "Analyze Problem," "Get Suggestions," and "Optimize Code." This menu sends requests directly to the OpenAI API for intelligent assistance.

• Tech Stack: JavaScript, OpenAI API, Chrome Extensions (MV3)

• Highlights: Integrated directly into LeetCode UI, saves time in competitive coding.

Hackathon-winning project integrating Federated Learning + Blockchain to securely manage blood donation records while preserving privacy.  

• Problem: How to use patient data from multiple hospitals to train a powerful AI model for risk assessment and donor matching without compromising patient privacy.

• Solution: A secure, multi-layered system that keeps data local to hospitals while allowing for collaborative AI training.

• Architecture Diagram: A system where multiple hospital nodes (each with local data and an AI model) communicate with a central blockchain. Encrypted model gradients are sent to a secure aggregator, which is part of the blockchain system. The final aggregated model is then sent back to the hospital nodes for local use, ensuring no raw data is ever shared.

• Tech Stack: PyTorch, Flask, TenSEAL (homomorphic encryption), Blockchain, Azure

• Highlights: Decentralized privacy-first health data system, scalable & secure.

IoT + ML system that monitors water quality and predicts usability using machine learning models.  

• Problem: Predicting the maintenance needs of an IoT-based greywater filtration system using sensor data.

• Solution: A machine learning pipeline that establishes a performance baseline for water quality classification and provides a clear mapping for future IoT sensor integration.

• Pipeline Diagram: Data from IoT sensors is collected and sent through a pipeline. The data is first pre-processed, then split into training and testing sets. The training data is used to train machine learning models (Logistic Regression and Random Forest), which are then evaluated on the testing data to produce performance metrics, visualizations, and a final summary.

• Tech Stack: Python, Random Forest, Logistic Regression, SHAP

• Highlights: Deployed with IoT sensors, interpretable ML with feature importance.

An intelligent classifier that automatically detects column types in CSVs and performs parsing + normalization.  

• Problem: Automatically categorizing and extracting structured information from a variety of messy, unstructured data.

• Solution: A robust two-stage pipeline that first classifies data using a high-performance ML model and then applies a specialized parser for extraction.

• Pipeline Diagram: A two-stage data processing pipeline. Stage 1 is "Data Classification" using a Random Forest model, which categorizes unstructured data. The classified data then proceeds to Stage 2, "Data Parsing," where it is processed by a semantic parser to extract structured information.

• Tech Stack: Python, Scikit-learn, Gemini API

• Highlights: 2-stage ML pipeline with accuracy evaluation and structured output.

A Java-based application for managing personal journals, including add, edit, delete, and view functionalities.  

• Problem: The need for a straightforward, local application for managing personal journal entries.

• Solution: A basic command-line or GUI application in Java that handles the core CRUD (Create, Read, Update, Delete) operations for journal entries.

• Visual Presentation: A user interacts with a simple interface (either a command-line interface or a graphical user interface) to perform actions on their journal entries, such as adding a new entry, editing an existing one, deleting an entry, or viewing all of their saved entries. The entries are stored persistently.

• Tech Stack: Java, Maven

• Highlights: CLI + desktop usage, persistent storage with clean design.

A fullstack web app for managing customer orders, invoices, and delivery tracking.  

• Problem: The need for a comprehensive system to manage the lifecycle of customer orders from placement to delivery.

• Solution: A fullstack web application with a user-friendly frontend and a robust backend that handles all order-related operations.

• Architecture Diagram: A user interacts with the React frontend, which communicates with the backend via REST APIs. The backend, built with Node.js and Express, processes requests and interacts with a MongoDB database for data storage. The system supports real-time updates and displays data in an intuitive interface.

• Tech Stack: React, Node.js, Express, MongoDB

• Highlights: Real-time order tracking, user-friendly interface, REST APIs.

2025 Contributions: 78
Repositories Contributed: 16+
Activity Breakdown: 🔹 100% Commits 🔹 0% Pull Requests 🔹 0% Issues 🔹 0% Code Reviews