Half a day of deploying, now about an hour
Deployments used to swallow most of a working day. I automated the whole journey, from building the code to checking its quality and shipping it, so it now takes 1–1.5 hours and nobody has to babysit it.
Hi, I'm Aarathy. I look after the systems behind enterprise software, so new features ship quickly, data stays safe and everything just keeps working.
I started out studying zoology. Today I look after a very different kind of ecosystem: the servers, pipelines and clusters a global business relies on every single day.
Deployments used to swallow most of a working day. I automated the whole journey, from building the code to checking its quality and shipping it, so it now takes 1–1.5 hours and nobody has to babysit it.
Every release came with the same long list of careful, repetitive steps. I built an AI-assisted workflow that works through them for us. That means about 80% less hands-on effort, and it still checks with a person before anything important happens.
The best outage is the one nobody ever notices. I set up alerts and dashboards that spot trouble early, like memory creeping past 90%, so we can fix it long before it reaches anyone using the service.
When something urgent goes wrong in production, I'm one of the first people on it. I get the service back, keep everyone in the loop, and then dig into why it happened, so the same problem doesn't come back.
I led the move of several environments, production included, onto a brand-new Amazon EKS platform. We found gaps in the new platform along the way, and I made sure they were fixed for every team, not just ours.
When critical vulnerabilities turned up in our containers, I moved the affected services onto safe, supported images. When a certificate authority was compromised, I removed it from every Java server that trusted it.
DevOps Engineer
Want to talk DevOps, cloud or keeping systems reliable? Or just say hello? Drop me an email or find me on LinkedIn. I'd love to hear from you.