Trained yolov5 to detect products on shelfs in stores. • Handling backend development with microservices kattna ml repo • • • Use Vgg16 as a feature extractor and cosine similarity to cluster the similar products. Use GANS for data augmentation Django, FastAPI, Celery, Redis. docker

Natural language processing models in Python to predict phishing content using TF-IDF features and Random Forest, Three models are in production. • Combined computer vision features with TF-IDF • Feature Engineering, Training and Testing of Models, Finding New Techniques, reporting Precision & Recall of models. Train Doc2vec on the phishing content to get the words Embedding, Python, Linux

Working on Object detection and tracking algorithms. Mapping of world coordinates to the 2D plane using homography. Training of InceptionV2 model, and Keras image classifier. Making of a real-time system for analytics. Performance improved from 5 fps to 33 fps. • Linux, Python, Multiprocessing, OpenCV, Tensor flow, TensorRT, Keras, Socket.io, NodeJs, and JavaScript