Skip to content

Projects

Key Projects & Applications

GitHub Projects

📝 Description:
• Collection of various data science and AI projects
• Includes web apps, chatbots, RAG-based applications, and AI agents
• Showcases work with Python, machine learning, and artificial intelligence
🔗 Link: GitHub Projects

HuggingFace Spaces

📝 Description: A collection of interactive AI and Data Science applications that includes various types of applications such as:
• Chatbots
• RAG-based systems (Retrieval-Augmented Generation)
• Streamlit applications
• Gradio applications

🤖 AI Agents: AI agents can be integrated into these applications to enhance interactivity and functionality. They can perform tasks such as:
• Answering user queries
• Providing recommendations
• Automating workflows

🔗 Link: Hugging Face

International Hackathon Projects

Voices of Intelligence: Lead with AI Agent Hackathon
• Built a multilingual voice summarization and translation pipeline using a modular agent architecture based on GenAI AgentOS Protocol.
• Integrated HuggingFace Transformers (BART, mBART), Google Text-to-Speech (gTTS), language detection, and Dockerized workflows for robust deployment.
• Enabled natural human-AI interaction through intelligent voice responses in multiple languages.
• Developed scalable AI agent solutions focused on seamless voice-based communication.
🔗Link: Lead with AI Agent Hackathon

National Hackathon Projects

CrisisPilot – Global Disaster Swift Response Assistant
• Developed a real-time disaster monitoring and alert system integrating Streamlit, Groq LLM, Serper API, and Discord Webhooks.
• Created a location-based alert pipeline for floods, earthquakes, and storms, enhanced with AI severity analysis.
• Implemented a 24/7 chatbot assistant to provide timely crisis information and support.
• Focused on rapid disaster response and improved public safety through AI technology.
🔗Link: Demo Link

Generative AI, RAG based Apps & Ai Agents

DocMind RAG System

📝 Description:
• Personalized Learning Paths: Creates Python, Data Science, and AI study plans based on user profiles including goals, knowledge level, and interests.
• Multi-Modal Interaction: Supports chat-based Q&A, quiz generation, and personalized study plan creation with adaptive explanations.
• Rich Learning Resources: Provide tutorials, documentation, and project ideas aligned with user preferences and progress.
• Session Persistence & Context: Maintains user session data for contextual conversations, tracking progress, and refining recommendations.
• Intuitive Multi-Tab UI: Clean Streamlit interface with dedicated tabs for Profile setup, Chat assistant, Resources, Practice quizzes, and Study plans.
🔗 Link: DocMind RAG System

RAG Document Assistant

📝 Description:
• Interactive Document Upload: Upload PDF or TXT files for seamless AI-powered document exploration.
• Retrieval-Augmented Generation (RAG): Uses semantic search with FAISS and Sentence Transformers to retrieve relevant text chunks for answering questions.
• Comprehensive AI Assistance: Supports Q&A, summarization, sentiment analysis, and adaptive responses strictly based on the uploaded document.
• Multilingual Translation & Audio: Translates answers into various languages and provides text-to-speech audio playback.
• Customizable AI Models & Debugging: Allows model selection and enables debug mode for detailed interaction insights.
🔗 Link: RAG Document Assistant

AI Teaching Assistant

📝 Description:
• Personalized Learning Paths: Customizes courses based on user profiles, including age, goals, and learning style, adapting recommendations as users progress.
• Structured Courses: Offers modular courses in Python, Data Science, and AI, designed for various skill levels with clear progression.
• Curated Resources: Provides high-quality external resources like tutorials and interactive courses that align with user interests.
• Hands-On Projects: Recommends project ideas suited to the user's learning stage, reinforcing skills through practical experience.
• Smart Chatbot: Engages users in personalized conversations, remembering their profiles and chat history for tailored responses.
🔗 Link: AI Teaching Assistant

Quiz Generator app

📝 Description:
• Quiz Generator app: Streamlit app that delivers timed multiple-choice quizzes related to Python, Data Science, and AI.
• Dynamic Question Selection: Randomly selects up to 30 topic-specific questions per session for varied practice.
• Multi-Domain Quiz Selection: Users choose a focus area (domain/expertise) from sidebar dropdown to start quiz.
• Interactive Timed Gameplay: Tracks score, enforces 60-second timers per question, and provides instant feedback on correctness.
• Progress Tracking: Allows users to monitor their quiz history and improvement over time.
• Link: Quiz Generator App

Machine Learning & NLP Projects

Amazon Reviews Classifier

📝 Description:
Amazon Reviews Classifier – TF-IDF & CountVectorizer with 8 ML models for classification.
• Compared Logistic Regression, SVM, Passive Aggressive Classifier, Random Forest, Naïve Bayes, Multilayer Perceptron, Gradient Boosting Classifier, and KNN models.
• Applied Random Under-Sampling to address class imbalance.
• Visualized class distribution and feature importance.
• Performed hyperparameter tuning to optimize model performance.
• Evaluated models using multiple metrics including accuracy and confusion matrix.
• Implemented text preprocessing steps such as tokenization, stopword removal, and lemmatization to enhance feature quality.
🔗 Link: Amazon's TopBooks Reviews: TfidfVec & CountVec

Toxic Mushrooms Classifier

📝 Description:
• Performed exploratory data analysis (EDA) to identify patterns, correlations, and class imbalance in mushroom species.
• Trained and evaluated LGBMClassifier, CatBoostClassifier, and XGBClassifier using AutoML with cross-validation for robust model selection.
• Tuned hyperparameters to optimize performance and reduce overfitting.
• Visualized feature importance for interpretability.
• Presented clear visualizations to explain model decisions, enhancing user trust and understanding.
🔗 Link: Revealing Toxic Mushrooms: Binary Prediction

Iris Flower Prediction App

📝 Description:
• Interactive Input Parameters: Users can adjust sliders for sepal length, sepal width, petal length, and petal width to customize their input for the prediction model.
• Real-Time Predictions: As users input their parameters, the app provides immediate predictions on the class of the iris flower based on the Random Forest Classifier model.
• Comprehensive Output: The app displays the predicted class along with the probabilities for each class, giving users a clear understanding of the model's confidence in its predictions.
• Educational Component: The app includes a section that lists class names with their corresponding index numbers, helping users learn about the different species of iris flowers.
• Model Training: The app utilizes the well-known Iris dataset to train the Random Forest Classifier, ensuring accurate predictions based on established data.
🔗 Link: Iris Flower Prediction App

Deep Learning & Time Series Forecasting

Tesla Stock Price Forecasting

📝 Description:
• Trained GRU and BiGRU models on Tesla stock price data for time series forecasting.
• Performed residual, seasonality, and trend analysis on opening and closing prices, including yearly breakdowns.
• Implemented early stopping to optimize model performance and prevent overfitting.
• Achieved 94% R² score on test, demonstrating model robustness for forecasting.
🔗 Link: Tesla Stock Secrets: Forecasting with GRU | BiGRU

Advanced Potato Disease Detection

📝 Description:
• Detected Early Blight and Late Blight in potato plants using EfficientNetB3.
• Applied data augmentation (rotation, zoom, flip) to improve generalization.
• Applied transfer learning with EfficientNetB3, achieving higher accuracy and rigorously evaluating model performance.
• Implemented image preprocessing techniques to enhance input quality and reduce noise.
• Developed a user-friendly interface for farmers to easily upload images and receive instant disease diagnosis.
🔗 Link: Advanced Potato Disease Detection | EfficientNetB3

Data Science & Exploratory Data Analysis

Gaming Evolution (1980–2023)

📝 Description:
• Conducted trend analysis using Seaborn and Plotly, performing comprehensive data cleaning and wrangling.
• Derived insights based on genre, views, reviews, and playtime.
• Formulated question–answer pairs from the findings.
• Visualized growth patterns and genre popularity over time to highlight industry shifts.
• Analyzed user engagement metrics to identify key factors influencing game success.
• Presented actionable recommendations for game developers based on data-driven trends.
🔗 Link: Gaming Evolution: Exploring Video Games (1980-2023)

Word Cloud Generator

📝 Description:
• File Upload & Text Processing: Supports uploading text, PDF, and Word files; extracts and processes text while removing common stopwords for meaningful word clouds.
• Customizable Word Clouds: Users can adjust size, word count, background, and contour settings to personalize their word cloud.
• Real-Time Visualization & Downloads: Displays word clouds instantly and allows downloading images (PNG, JPEG, SVG, PDF) and word count data as CSV.
• Word Count Table & Interactive Interface: Shows detailed word frequency tables with an easy-to-use Streamlit layout and sidebar customization.
• Educational & Open Source: Acts as a learning tool for text analysis and visualization, with open-source code for community collaboration.
🔗 Link: Word Cloud Generator Streamlit App

Web Scraping & Data Collection

OLX Scooty & Scooter | Web-Scrapping

  • Description: Perform web scraping using Beautiful Soup & Requests library to extract product data.
  • Collected and cleaned detailed listings including prices, locations, time and title.
  • This helps in Analyzing market trends and pricing patterns based on the scraped data.
    🔗 Link: OLX Scooty & Scooter | Web-Scrapping

Ebook for Kids

  • Description: Created an interactive Ebook for Kids introducing A-Z alphabets themed around AI & Data Science concepts.
  • Designed colorful, engaging visuals and simple explanations to make complex topics accessible for young learners.
  • Aimed at inspiring early interest in technology and data literacy from a young age.
    🔗 Link: Ebook for Kids