Rockbot-AI: Comprehensive Multi-Agent Assistant

Rockbot-AI is an advanced AI assistant built for autonomous task execution and complex interactions. Its core strength lies in its Multi-Agent Architecture, which utilizes specialized agents (General, Translator, Creative Writer, Problem Solver, Task Executor) to efficiently handle Q&A, content generation, and step-by-step task planning.

The platform supports multi-modal processing (image,pdf,doc etc upload and analysis via OpenAI Vision API) and features adaptive learning for context retention. User experience is streamlined through a modern, responsive React-based UI that enables conversation saving, sharing, and essential document export capabilities like PDF, Doc, and text file generation.

Tech Used:JavaScript, Python React, Shell Scripting, LLMs, Multi-Agent Frameworks, OpenAI Vision API,OpenRouter, SQLAlchemy, Tailwind CSS and others

Assistant-using-RAG-with-Cohere-API

This project implements a Retrieval-Augmented Generation (RAG) pipeline using the LangChain framework and Cohere API. It allows users to ask questions about real estate properties (e.g., "What is the rental yield of Greenview Estate?"), and answers are generated based on document data stored in a local directory.

Tech Used: Retrieval-Augmented Generation (RAG), LangChain, Python, Cohere API, FAISS Vector, LLM, Dotenv

Lead Qualifier with Airtable Integration

Developed an AI-driven lead classification and entity extraction tool for real estate workflows. Used Cohere’s LLMs to extract structured information (location, intent, budget) from free-text messages. The processed leads are automatically saved to CSV and pushed into Airtable for workflow automation.

Tech Used: Python, Cohere API, Airtable API, NLP, Entity Extraction, CSV Automation

AttendSure: AI-Powered Attendance

Developed an intelligent attendance system using facial recognition and computer vision. Leveraged transfer learning (ResNet) for face detection and engagement analytics for classroom insights.

Tech Used: Python, OpenCV, Firebase Realtime Database, Face Recognition Library, Flask, HTML, CSS, JavaScript, Bootstrap, Barcode Scanner Integration, Real-Time Data Visualization

CityPulse: Smart Urban Issue Reporting Platform

CityPulse is a full-stack urban issue reporting platform that empowers citizens to report real-world city problems like potholes, broken streetlights, and garbage dumps directly from their location.

Tech Used: JavaScript, React.js, MongoDB, OpenStreetMap API, REST API, Axios, Tailwind CSS, Render

ChatBot for School Management System

Developed an AI-powered chatbot to automate student and staff queries within a school management system.

Tech Used: Python, PHP, MySQL, Natural Language Processing (NLP), Flask, HTML, CSS, JavaScript, AJAX

Birds Classification using Transfer Learning

Developed a deep learning model to classify multiple bird species using advanced transfer learning techniques with 98.47% accuracy.

Tech Used: Python, TensorFlow, Keras, Transfer Learning, EfficientNetB3, ResNet152V2, Matplotlib, Scikit-learn