In this project I created a responsive frontend design created using HTML,CSS and Bootstrap 4 for a fictional app called tindog (Tinder for Dogs).

In this project we made an application to recognize the American Sign Language using OpenCV and Convolutional Neural Networks. Created my own dataset of 19200 images for training the network on google Colab.

In this project, I worked on real-time proprietary data for Bertelsmann partners AZ Direct and Arvato Finance Solution to identify customer segments that form the core customer base of the company using unsupervised learning techniques. Data cleaning was an integral part of the project since the data used here was real-world. Techniques like Principal Component Analysis were also used for Dimensionality Reduction.

Trained an image classifier to identify a total of 102 flower species. Data Augmentation was used to bring variety in the dataset. I also made a command-line interface for training and testing our model with various parameters using the ArgumentParser library in Python. Transfer Learning with VGG16 and Densenet121 was used to train our neural network.