At Zycus, the company was growing fast and the platform was growing with it. I was tasked with starting a DevOps implementation. We started with hiring for the team, working on a roadmap, working with external consultants in areas where we were venturing new and creating an implementation plan based on our maturity. We eventually created a team of about 7 members.Our goal was to increase our DevOps maturity with every release. This included CICD implementation, bringing in the right monitoring tools for the problems to solve, etc.We aimed for immutable deployments and tech stack wise, we went with Python, Ansible, Jenkins stack for CICD. For monitoring we wrote a self-configuring Icinga2 implementation, for other maintenance tasks we went with Ansible Tower.

At BrowserStack I was part of the platform team. Here I got to work in a fast paced, hyper-focused environment where everyday bought new challenges. In my stint here, I coded on Ruby to automate platform tasks. We worked on AWS and VMware tools to automate provisioning, deployments, maintenance, etc. We wrote platform tests in Rspec.

Since I knew Java, I began contributing to the product code of one of our applications. This was Servlets-JSP and Oracle, MySQL and MSSQL. This was an internal ML tool where we looked at datasets, trained about 4 models and gave out results for multiple clients.This was a time when we started increasing our footprint in terms of platform size and it was evident that we now needed to standardize and automate. It was clear that for the platform automation, Python would be a batter suited tool of choice and I started porting all automation to Python. In a few months, all our automation was in Python and code base was growing and we on-boarded one more team member to contribute.Together we built a new internal automation tool that automated even more tasks, including deployments with a real-time updating UI with nodejs.It involved a lot of standardization to be done, both from the product teams and the platform team. We went from 30 manual deployments/week to 200+ automated deployments/week in this phase limited only by product release speed.

Started off as an application support engineer where my job was to do the deployments on monolithic Java applications, debug problems and take care of monitoring alerts. I was trained in Java so about 6 months into the job, I started automating a bunch of tasks I did and that became the theme for next tenure as a support engineer. I wrote an internal support application to take care of tasks like application restarts, maintenance tasks, etc. This was done in JSF :D