Necati Demir, PhD

An end-to-end CI/CD system I built for a fintech startup.

I used GitHub Actions to run Python tests, build Docker images, and push the images to the AWS container registry. I also used Terraform to create the AWS infrastructure and Kubernetes and Helm for deployment.

Built the data ingestion and machine learning system's infrastructure using Kubernetes on GCP. I also worked as a back-end developer and manager on this project.

I worked as a back-end developer implementing the API with Python and MongoDB and as a DevOps engineer implementing the CI/CD system with GitLab, Helm, Kubernetes, and Docker. I also managed a team of back- and front-end developers using the Kanban methodology.

This project's most significant accomplishment is the DevOps part, which dramatically improved the product's fate. After the changes, the product's NPS changed from negative scores to between 90-100, and the average number of bugs decreased from 300 in production to less than 10. The details of the DevOps cycle that I designed and implemented can be found here: https://gist.github.com/ndemir/48caa5399f1bfe38bd6939f5d32f2c96

I built the infrastructure and architecture of a small telecommunication company and also worked as a back-end developer and manager on this project.

RESPONSIBILITIES

• Took a role as a back-end developer
• Handled the implementation of API (Java, MySQL)
• Contributed to the implementation of the voice call and IVR systems (FreeSWITCH, Asterisk)
• Worked on the implementation of bulk email systems
• Took the role of a system administrator and built a cluster of email servers to handle load balancing (HAProxy, Postfix)
• Managed a team of back-end and front-end developers

Switched to AWS Step Functions as a solution for the company, which had a dozen Lambda services that talk to each other and trigger the next from the previous one in a chain model without any centralization.

I refactored the existing codebase to make it compatible with the new architecture. I also implemented the Terraform components to spin up the Step and Lambda Functions.

A project for a software analytics platform company, which involved designing a system to fetch information on the libraries used in an application.

I led a team of developers and DevOps engineers. I designed the software architecture, defined the standardization, and did the initial implementation. I also took the role of a DevOps engineer and wrote Terraform and converted those to Amazon ECS tasks.

SIG is a leading finance company in the United States, processing terabytes of transactional data—including buy/sell options and more. Various analysis and production issues resulted in incorrect reporting of some field-level data. I was responsible for identifying and correcting these errors by processing large volumes of data using PySpark.