Asjad Haq Email and Phone Number

• Developed and implemented Azure DevOps pipelines for CI/CD processes, automating infrastructure builds, testing, and deployment workflows.• Used Terraform to develop re-usable modules that encapsulates, standardized and secured infrastructure to enable business partners to quickly deploy solutions.• Implemented Well-Architected-Framework for azure cost optimization saving 10% of cloud costs.• Used Azure SDK for python to process data from azure resource manager API's to determine the usage of resources and alert accordingly to reduce costs.• Applied Azure Center for Internet Security policies to increase secure score from 60% to 85%.• Improved build agent and deployment architecture by migrating the agents to Azure Kubernetes Services which improved security posture and reduced costs

Worked on the creation of two types of recommendation (visual similarity and personalized recommendation for customers) for magento (an e-commerce platform developed by Adobe) and setting up the generic data pipeline to be used by multiple magento clients.Understood the relational EAV structure of the database for a magento client. First ran the queries on the DB to extract the required data for a client and then shifted to using the magento REST API's to extract the data and pre-process in the required form for the algorithms to run, and also wrote unit tests.Setup the recommendation endpoint API using the Flask web framework on an AWS EC2 server by configuring nginx and the virtual environment and establishing a service to be used on the clients website and tested the endpoints using Postman.Visual similarity was done using transfer learning and changing a few layers of the neural network. Explored avenues on how to improve the feature extraction process of the neural network like using the YOLO algorithm for object detection in an image to get better results.Worked on writing the Collaborative filtering algorithm which was used to generate personalized recommendations for a particular user. Extracted the data and processed it into required form for content based filtering which included extracting textual data to be used by the NLP models. Python was used in setting up the whole pipeline with the use of popular libraries like pandas, numpy, cv2,PIL, tensorflow, sklearn e.t.c.

Used powerBI to create dashboards and reports by pulling the data from the SQL server and excel files which were used to gain buisness insights.