Apr 2021 – Jan 2023 · 1 yr 9 mos

• ⦿ Took the initiative and developed a tool in-house called "Atlas" that automates setting up Load balancers and related resources in AWS, including (Nginx) proxies, using a Web GUI with Terraform, and Golang as core backend technologies. This formed the base for rolling out 40+ different lines of testing in BigBasket's non-production environments. (Blogpost: https://tech.bigbasket.com/atlas-streamlining-bigbaskets-40-lines-of-testing-across-80-microservices-in-non-production-459040947519)
• ⦿ Served as a liaison between Platform and DevOps teams to effectively work on certain strategic initiatives aimed at Cloud-Infrastructure.
• ⦿ Brought more visibility into our Aerospike clusters by integrating them with Prometheus.
• ⦿ Took the initiative and proposed the practice of spinning Newrelic dashboards, and alerts through Terraform. This also avoids duplicating Newrelic dashboards, and alerts, while maintaining consistency, even when microservices grow from 10 to 100+.
• ⦿ Introduced Github's Gh-ost schema migration tool (https://github.com/github/gh-ost) to Bigbasket and practically eliminated the need for taking late-night (db related) scheduled downtimes every few months. Made the Engineering team's life easy. Gh-ost rocks!

May 2018 – Mar 2021 · 2 yrs 10 mos

• ⦿ Became the go-to person in BigBasket for all things regarding Kubernetes.
• ⦿ Took the initiative and developed an in-house tool called Heva(https://tech.bigbasket.com/heva-achieving-consistency-in-microservice-deployments-through-manifest-discipline-c3d560525781) in Go/Golang for validating and merging Helm Values files. This brought a common standard (and even auto-baked production learnings) for all microservices' configurations in Bigbasket.
• ⦿ Took the initiative and came up with an initial implementation splitting Kubernetes deployments so that half the pods run on spot vs on-demand. This has tremendous cost benefits.
• ⦿ Gave a talk on "Scalability and Reliability in BigBasket" at Large Scale Production Engineering - India (LSPE-IN) meetup. (Slides below)
• ⦿ Gave numerous techtalks to teams regarding Helm, Kubernetes, Docker, and Istio and inspired them.
• ⦿ Lead various microservices migration to K8s in production. This increased agility while reducing costs.
• ⦿ Laid ground rules for launching the initial set of microservices, that became a reference for many more in the longer run.
• ⦿ Devised a strategy to reduce Kafka connections count from 1 lakh to 10,000 per each Kafka box.
• ⦿ Inspired many application teams to treat Reliability, Monitoring as one of the foundational principles for developing microservices.
• ⦿ Prepared numerous templates for BigBasket's Engineering team, including Incident Post-mortem (R.C.A) templates, Tech spec templates, and Blogpost templates. These templates allow developers to have creative freedom while maintaining structure and discipline in their writing.
• ⦿ Lead numerous production cutovers that involved Mysql databases, Kafka, Aerospike upgrades, and service migrations between multiple K8s clusters.
• ⦿ Mandated and lead Infrastructure As Code (I.A.C) practices in our team to spin infrastructure on the AWS cloud.
• ⦿ Wrote many technical specs/docs/wikis for all the things that are deeply technical and inspired many teams to do the same.

Apr 2016 – Apr 2018 · 2 yrs

• ⦿ Worked directly with CTO, lead a team focused purely on Mysql database performance, shaved millions of SQL queries hitting a monolith database, and improved the response time of bigbasket.com tremendously.
• ⦿ Authored a Kafka Client in Python, built on top of confluent-kafka-python library(https://github.com/confluentinc/confluent-kafka-python/), but has pretty much all goodness of Celery(https://docs.celeryproject.org/en/stable/). (Handles delayed processing, duplicate processing, celery style decorators for the producer, and consumer logic.) This has been running great in production for 6 years, processing over 16 million msgs every HOUR!!
• ⦿ Planned and lead numerous production cutovers, which involve database upgrades, and migrations.

Jun 2014 – Mar 2016 · 1 yr 9 mos

• ⦿ Developed the UI layer(Websockets + AngularJS) and web service layer(Pyramid+RabbitMQ) to handle (Jiva)version agnostic clinical content.
• ⦿ Automated the build and release management and lead the team to handle releases efficiently.

Jun 2012 – Jun 2014 · 2 yrs

• ⦿ Lead the Continuous Integration Build and Release Management using Jenkins. Automated many manual procedures which were killing developer time and productivity.
• ⦿ Introduced the concept of Techtalks & Code-athons to the company and successfully organized two code-athons.

Oct 2009 – Jun 2012 · 2 yrs 8 mos

• ⦿ Developing various configurable features in Jiva which behave differently across different Clients.
• ⦿ Developing a suite of scripts for auditing any developer's code based on certain Jiva specific standards.
• ⦿ Learning & training many people on the ins & outs of Python Packaging tools like Buildout, Pip(in Virtualenv). Contributed lot of improvements to the build tools that enhance developer productivity inside the company.
• ⦿ Working on an In-House developed framework to handle loading client data(in CSVs) into Jiva’s database, using ZeroMQ.
• ⦿ Developing in-house python scripts to automate the Administration part of Github Teams, Repos. Using Github API V3. Also interfacing with LDAP/AD system.

May 2007 – Sep 2009 · 2 yrs 4 mos · Greater Hyderabad Area

• As a Software Developer,
• ⦿ Developed the whole application using Python/Django, designed UI screens, designed database schemas.
• ⦿ Lead A/B Testing exercises.
• ⦿ Built Payment Gateway Integrations.
• ⦿ Built an SMTP server to host all the emails corresponding to Peoplelex.com website users

Feb 2006 – Apr 2007 · 1 yr 2 mos · Greater Bengaluru Area

• I used to be part of PLM (Product Life Cycle Management) project which mainly deals with enhancing the CNH(Case New Holland) Ematrix application from the version 10.0.0.5 to the latest version 10.6SP2. This includes developing and testing the new features like Task Delegation, Type Ahead, Modified ECR-ECO process, tightly-coupled Route Board processes. Enhancements at the schema level as well at the UI level are done, to ensure it supports some distinct features of previous version too.