1. Project had multiple objectives in addition tothe major objective of detecting set of objects and applying OCR on top of it to read the digits from images/videos

2. Tracking of an detetced objects to keep the unique count of objects passing through camera frame

3. Capturing the direction of an object passing through camera frame.

4. Used YOLO(v3) for object detection, easyocr & tesseract for OCR, deepsort for object tracking and statistics & geometry concepts to make logic more reliable.

Training state of the art customised object detection model on customer data and extract useful information from the objects being detected.

Working on to understand customer cloud usage and help them recommend cloud services over existing services that can help customers save on their billing cost.

Also predicting the Virtual Machine status whether it is idle or not based on user's virtual machine usage data

Trained RNN to predict the amazon reviews as positive or negative towards product purchased by customers by using text review given by customers.

Use Customer Reviews (text) to predict whether the review is positive / negative towards the product Used Data sets from kaggle to classify the reviews. Data consisted of text review and the ratings (1 to 5) Worked on Python, Machine Learning, Jupyter Notebook, Colab

Whether the two questions are similar to each other or not by using Question(text) Data Worked on Python, Machine Learning, NLP, Jupyter Notebook, Colab

Used Network Data to suggest a particular friend to user Worked on Python, Machine Learning

Used sensor dataset to predict human activities for example walking, sitting, sleeping. Worked on Python, Deep Learning, Jupyter Notebook, Colab

Trained machine learning algorithm to predict the winner of a cricket match Used supervised machine learning including regression and classification techniques Worked on Juptyer Notebook, Python, Machine Learning, Data Analysis

Designed a prototype that can scan a drivers license to unlock the vehicle and also same can be used at traffic signal to detect the traffic signal rule breakers Hardware sensor based Developed prototype Used Arduino Programming