BTech CSE student with hands-on experience in Python development, machine learning, and web APIs. Built and deployed multiple end-to-end ML projects involving real-world data, model deployment, and frontend-backend integration. Passionate about solving real problems through intelligent systems.
• Full-stack web app for handwritten digit recognition using a custom CNN (98.9% accuracy on MNIST).
• Users can draw digits on a canvas or upload images; app predicts the digit and shows model confidence in real time.
• Features intelligent image preprocessing and a clean, responsive frontend. Built with TensorFlow, Flask, and vanilla JS/HTML/CSS.
• Fashion image classifier trained on Fashion MNIST using a CNN with Batch Normalization and Dropout (~91% test accuracy).
• Flask API for image uploads and predictions; minimal frontend for easy testing; robust preprocessing for consistency.
• Focused on end-to-end workflow: model design, optimization, deployment, and frontend integration.
• Real-time waste classification web app using custom-trained MobileNetV2 model.
• Identifies 9 types of waste with confidence scores and disposal tips.
• Built with Flask, TensorFlow, and responsive frontend (HTML/CSS/JS).
• Web-based tool that detects bias in team chat messages in real-time.
• Identifies gender-coded words, microaggressions, and offers inclusive rewrites.
• GNEC Hackathon Project for AI and Gender Equality (SDG 5).
• Web-based platform for skill exchange among university students using Time Credits (TCs).
• Features include secure transactions, dynamic request cards, and user-friendly dashboard.
• Built with vanilla HTML, CSS, and JavaScript for optimal performance.
• Built a tone-detection app using TF-IDF with Naive Bayes.
• Real-time feedback on email tones (Friendly, Formal, Aggressive, etc.) via Flask API and JS frontend.
• Developed a Logistic Regression model to detect fraudulent job listings.
• Deployed with a clean, responsive frontend using Flask and vanilla JS.
• Trained a Linear Regression model with feature engineering and Grid Search.
• Built a deployment-ready interface to predict mileage from user inputs.
• Used Librosa to extract audio features and predict genres via Flask API.
• Handled MP3/WAV input formats and resolved backend deployment issues (CORS, NumPy compatibility).
• Created a content-based recommender system in Python.
• Suggests similar movies based on title and genre correlations.
Lovely Professional University — 2nd Year