Aleksandar Radonic

Used Python to extract data from an AWS S3 bucket, transform the data using Python, and load it into an RDS PostgreSQL database table. The credentials for the database were stored in AWS Secrets Manager. The project was wrapped in a Docker container, could be deployed to an AWS ECS instance, and included AWS CloudWatch logging. The spooler was automatically executed when a file was uploaded to S3.

Deleted old BI dashboards and recreated them based on new and updated data sources. The dashboards were written in PostgreSQL and sometimes the output was augmented using Python. Using the BI tools API system, I also restructured the permission system for who could see which dashboards and what they were allowed to do in them.

Completed work toward architecting and building a new analytics database that will be a clone of an existing database along with added columns and a new structure to suit the needs of the analytics department. The database is based on data delivered daily in S3 and data delivered continuously through SQS. The data is spooled using ECS Python tasks. I set up the RDS database and coded and deployed the ECS Python tasks, and I am responsible for maintenance.

Due to an acquisition of the company I worked in, a data feed from an external company had to be created. I created this using a Python Lambda API endpoint that writes to an SQS queue. The external company can send messages to our SQS queue through the endpoint, and then an ECS Python spooler task unpacks and spools the messages into our RDS database. I set up the Lambda endpoint and the SQS queue, created the table structure and added tables to the RDS database, and coded and deployed the ECS Python tasks. I am also responsible for maintenance.

To understand the theoretical foundation of convolutional neural networks, I gave myself a mission to understand it so well that I could hardcode it in NumPy and make it return my voice. I succeeded both in understanding the theory and hardcoding. Next, I wanted to hardcode it into an iPhone app to see if an iPhone was powerful enough to train a network, so I hardcoded the network in Swift.

I developed this platform for a client. It had an admin page used for creating and changing online coaching sessions and a front end for customers to view and follow different courses. I also integrated Stripe as a payment gateway. In this full-stack assignment, I created the back end in PHP and MySQL and the front end in HTML, JavaScript, and CSS.