Robust AI Model: Trained on a large dataset of labeled microscopic cell images, achieving a loss of 0.2286, accuracy of 0.9418, validation loss of 0.4207, and validation accuracy of 0.8856.

Mobile Application: Developed with Flutter, this app enables healthcare providers to upload bone marrow cell images for real-time classification.

Tools & Techniques: Google Colab, Python, CNN, MobileNet, Classification, Kaggle, Flutter, Keras, TensorFlow, Firebase, Flask API, GitHub.

Project Achievements: Finalist in the Google Solution Challenge 2024, 1st place in the Build with AI Idea Competition at GDSC CUI ATD, Runner-Up in the SparkPitch Idea Competition at Tech Vision, Islamia College.

Developed a multi-task neural network model for object detection and classification.

Implemented custom data generators to handle image augmentation and preprocessing.

Achieved an overall classification accuracy of 87.08% on the test set.

Utilized Mean Squared Error (MSE) for bounding box regression and categorical cross-entropy for classification.

Employed advanced techniques such as early stopping and learning rate reduction to enhance model performance.

Developed a cross-platform desktop application using Flutter to integrate the trained model for real-time gesture-based media control.

Tools/Techniques: Python, TensorFlow, Keras, OpenCV, ImageDataGenerator, NumPy, JSON, FlaskAPI, Flutter.

Integrated a U-Net for precise segmentation of radiographic images and a MobileNet-based model for classifying osteoarthritis severity into five categories: Normal, Doubtful, Mild, Moderate, and Severe.

Implemented and compared various techniques using Deep Learning and traditional Machine Learning methods to evaluate performance.

Applied early stopping and learning rate reduction techniques to enhance model training and prevent overfitting.

Tools/Techniques: Python, TensorFlow, Keras, OpenCV, U-Net, MobileNet, CNN, ImageDataGenerator, NumPy, JSON, FlaskAPI.

Implemented various classification models (Logistic Regression, SVC, Random Forest, and LGBM) to predict vehicle maintenance needs based on real-time data from an OBDII reader.

Optimized model performance using specific hyperparameters, including penalty and C for Logistic Regression, kernel and C for SVC, and various parameters for Random Forest and LGBM.

Achieved high accuracy with the LGBM model (87.6%) and evaluated performance using metrics such as accuracy, precision, recall, F1-score, confusion matrix, mean squared error, mean absolute error, and R² score.

Integrated real-time vehicle data from the OBDII reader into a FlaskAPI server, which processes the data and makes predictions on vehicle maintenance needs.

Developed a mobile app that allows users to view real-time predictions and vehicle data, providing an accessible interface for users to monitor their vehicle's condition.

Implemented key functionalities including user authentication, service management, search and booking capabilities, payment processing, and notifications.

Designed the app with future enhancements in mind, such as adopting Clean Architecture, integrating secure data handling practices, and enabling dynamic QR-based bookings.

Tools & Techniques: Flutter for cross-platform development, GetX for state management, Firebase for backend services and authentication, Stripe for payment processing, and QR code generation libraries for dynamic bookings.

Created an e-commerce app focused on electronics devices with a fully responsive UI.

Utilized GetX for state management and navigation to ensure a smooth user experience.

Designed the app to be cross-platform, accommodating various devices and screen sizes.

Implemented Firebase for user authentication to securely manage user accounts.

Integrated APIs for efficient data handling related to electronic products and their management.

Led the design and development of a user-friendly interface, focusing on a seamless user experience.

Implemented reusable UI components to ensure consistency and efficiency across the platform.

Tools & Techniques: Utilized Flutter for cross-platform development, adhering to responsive design principles for maintainability and scalability.

Challenges Overcome: Successfully addressed complex UX challenges such as real-time booking updates and optimizing performance for mobile users.

Outcome: Delivered the app on time, which resulted in a 20% increase in user engagement within the first three months post-launch.

Implemented state management with GetX for efficient data handling and a seamless user experience.

Designed the app to facilitate easy complaint submission and tracking, ensuring effective communication between residents and management.

Tools & Techniques: Flutter for cross-platform development, GetX for state management, Firebase for backend services.

Tools & Techniques: Firebase Authentication, Firestore Database, Firebase Cloud Messaging, Flutter Local Notifications.

Created a Car Wash App where users can book services and visit the garage.

Used Flutter for cross-platform development to reach a broad audience.

Implemented GetX for efficient state management and a smooth user experience.

Utilized Firebase for user authentication and data storage, ensuring secure and reliable backend services.

Developed a TikTok-like app with Flutter, providing a video-sharing experience across platforms.

Utilized Firebase for database management to handle video storage and user data.

Implemented GetX for state management to ensure smooth data handling and real-time updates.

Designed a responsive UI that adapts to various screen sizes, enhancing cross-platform usability.