Projects
AbroadStudy.AI
Retrieved Universities’ data from Kaggle and tuned a classification to predict the chances of getting admission and used different libraries. Used Natural Language Processing (NLP) and NLTK which recommend other universities to get admission as per students' profile.
Bright Look
Collect images of ears and tuned multiple CNN models and used transfer learning techniques to detect ears for wearing earring with an average of 90% accuracy for each case. Used DeepSORT and StrongSORT for live tracking of ears and deployed a FastAPI webapp for increased usability for local customers.
Quest Cricket Player
Used Labelbox for annotate sports person classification and tried YOLO and MaskRCNN with the format of COCO and JSON. Tuned hyperparameters of each individual model by examining results leading improved the accuracy of trained model by at least 5%. Used ML model for identifying multiple players.
