My Projects
Explore my work in AI and machine learning. Each project solves a unique problem using cutting-edge technologies.
YT Sentiment Analysis
Analyzed YouTube comments using sentiment analysis models and deployed the system with an end-to-end MLOps workflow.
Plant Disease Classification
A CNN-based system that identifies plant diseases from leaf images with 95% accuracy. The model was trained on a dataset of 50,000+ images across 38 disease categories. The system helps farmers detect early signs of plant diseases, potentially saving crops and increasing yields.
Vehicle Detection System
YOLO-based real-time vehicle tracking and detection system for traffic analysis. The model achieves 89% mAP on the COCO dataset and processes 30 FPS on consumer hardware. Potential applications include traffic monitoring, parking management, and autonomous vehicle systems.
Image Captioning System
An deep learning model that generates human-like captions for images by understanding visual content.The system uses a CNN (ResNet) for feature extraction and an LSTM-based decoder with attention mechanism for natural language generation. Trained on the MS COCO dataset with over 80,000 images and corresponding captions, it helps in tasks like accessibility, content automation, and image indexing.
Movie Recommendation System
A content-based recommender that suggests similar movies using cosine similarity on feature-engineered metadata. Built with a dataset of 10,000+ movies, the system uses scikit-learn's cosine_similarity to calculate similarity scores and return the top 5 closest matches. The app is deployed using Streamlit for a fast and intuitive user experience.
Job Scraper
An intelligent job recommendation system that analyzes your resume using Gemini AI and fetches real-time job listings from LinkedIn and Naukri via Apify actors. The app extracts and summarizes PDF resumes, identifies skill gaps, builds a learning roadmap, and uses Gemini to generate keyword-based queries for targeted job scraping—all inside a sleek Streamlit interface.
House Price Prediction
A regression-based ML model that predicts house prices based on features like location, area, and number of rooms.Trained on a real estate dataset using linear regression and ensemble methods to achieve high accuracy in price estimation.
Virtual Voice Assistant - Jarvis
A Python-based voice assistant that responds to user commands using speech recognition and text-to-speech.It can answer queries, fetch news, generate responses using OpenAI API, and perform tasks like playing music or reporting time—creating a hands-free smart assistant experience.
Toxic Comment Classification
A multi-label text classification system that detects toxic language, threats, and hate speech in online comments. Built using the Jigsaw Kaggle dataset, the pipeline includes TF-IDF vectorization and models like Logistic Regression and Naive Bayes for robust toxicity detection.
Weather Forecast
A responsive web application that displays real-time weather data for any city using the OpenWeather API. Built with HTML, CSS, and JavaScript, it shows temperature, humidity, weather conditions, and icons in a clean UI.