1. Responsible for training the CNN models for both regression and classification with 2D annotation datasets and analyze the loss and accuracy for the same.

2. Worked on implementing the undersampling approach for reducing the class imbalance problem in the dataset.

3. Responsible for validating the annotations provided by vendors by drawing bounding boxes using python scripts.

4. Wrote python scripts for -

a) preprocessing of the datasets to improve accuracy for the models.

b) organizing the datasets with different pixel sizes for better analysis of the input provided to the ML models and how they perform.

5. Implemented algorithms to check the blurriness quotient of images with change in camera focus.

Project Name: JD WILLIAMS, N Brown Group [2016 - Present]

Role: Hybris Developer

Description: A Global Multi-Channel eCommerce Transformation journey for a catalog retailer based in United Kingdom. The goal is to consolidate and transform the existing websites into SAP Hybris enabled eCommerce platform supported by Oracle Retail Merchandising Suite of products, Oracle Fusion Middle ware and existing legacy solutions.

Roles and Responsibilities:

1. Played the role of an Agile Engineer helping my team to develop eCommerce Stories for the key websites through the SAP Hybris platform.

2. Received good exposure and working experience to manage the quality of the deliverable

through customizing the Hybris out-of-the-box functionalities, writing of Junit tests and validating the outcome through the SonarQube reports.

3. Have extensively worked in JIRA and Confluence that has helped both in planning and excellent collaboration between a distributed agile team working in Continuous Integration (CI) and deployment (CD).