Jan 2023 – Mar 2024 · 1 yr 2 mos · Bengaluru, Karnataka, India · Remote

• At Hubilo, me and my team builds and maintains microservices and serverless functions that are responsible for critical business functions. This includes payment systems, lead generation and notification pipelines. We primarily use Node.js with TypeScript and there's a mix of NoSQL/SQL databases along with an event driven architecture based on Kafka. The services cater to millions of requests per day.
• A few things I built here:
• Attendee leaderboard - We built a leaderboard that categorises attendees based on thousands of attendee engagements happening per second. I lead this project from scratch. Its architecture is based on AWS Fargate, Apache Kafka and Redis. I also devised the most optimal algorithm we could use and the scoring happens within a few seconds for thousands of attendees. This feature continues to be a pivotal piece in the Hubilo Webinar legacy.
• URL shortener - I authored and designed a URL shortening service which is capable of shortening thousands of URLs per second. Here too, I devised the most optimal hashing algorithm for this use case and with efficient use of distributed caches and edge functions, the response time stayed under 50ms for users around the globe. This was my solo venture and also won me the Hubilo Limitless award.
• Data archival - I designed a pipeline by which we could offload historical data from our main databases (MySQL, MongoDB, Aurora) into cold storage (AWS Glacier). I laid the foundations for this framework and lead my team to its delivery. Terabytes of data can be moved without burdening any service or database with the help of intelligent data batching and streaming.
• Telemetry - I led development for a framework based on Prometheus and Grafana that could easily be incorporated into our services. It allowed us to measure metrics from VMs and ephemeral services like Lambda functions. We are now more proactive as all anomalies/p0s are reported before the customers encounter them.

Jul 2021 – Dec 2022 · 1 yr 5 mos · Bengaluru, Karnataka, India · Remote

• A few other things I did at Hubilo:
• Submeta - The only metaverse related feature in Hubilo was built by our team. It is very realtime with micro interactions taking place every second. We had to make sure we correctly track the position of the user with respect to other users on the screen. I lead this feature from scratch. We are able to correctly serve 30K requests per second. Data consistency along with high throughput is maintained by efficient use of Lua scripts that are run in a Redis cluster. Realtime interactions are pushed via persistent Websocket connections and different services talk to each other via an event driven architecture based on Kafka.
• Scaling our realtime servers and making them robust - I was able to re architecture our Websocket server cluster and collaborate with other stakeholders in making it more performant and robust to race conditions. We were able to eliminate race conditions around data consistency while still supporting 10K open Websocket connections per node.
• User RBAC - I guided my team to build an in-house RBAC service that provides access control to various resources in an internal company dashboard. Each user is associated to one or more roles with each role having its own permission matrix. The service is the sole owner of its cache and database and all the other services delegate RBAC requirements to this service. High availability is guaranteed by load balancing requests across a cluster.
• Testing culture and framework - I worked on laying the foundation for a testing framework by which we could achieve more than 90% test coverage for the services we built. The automation and coverage around this ensured we shipped robust features to production with very minimal or in some cases even 0 bugs during the QA/release cycle. Our team holds the record for the maximum amount of test coverage for any service across Hubilo!

Aug 2021 – Mar 2022 · 7 mos

• I collaborated with the engineering team at Musalleen to solve their pain points. This role was a mix of things where I provided insights and advice on how to strengthen the overall engineering process and debugged hard to solve issues.
• I frequently did code reviews and as well as coached my peers.

Dec 2018 – Jul 2021 · 2 yrs 7 mos

• At Musalleen, I was involved with building responsive web apps in JavaScript and React.js and lending a hand to a Python Flask RESTful backend. I was also involved with Android apps in Kotlin, hybrid apps in Dart and Flutter, and occasional DevOps. The role here was almost full stack and what I got here was being a ninja.
• I and my team worked as much as agile as possible while keeping code quality paramount.
• Deployment for APIs and servers were done either on Heroku or inside scalable Docker containers orchestrated by AWS ECS. Sometimes we dived into serverless too. Data was relational and was stored inside SQL databases or AWS RDS instances.
• For continuous integration and deployment, I and my team set up a Jenkins server on AWS. This helped reduce the feedback loop in our APIs, web, and mobile applications since all the testing and deployment was now automated.
• Static asset delivery and hosting were done via S3 and CloudFront.
• Occasionally, we did make use of Terraform to automate the provisioning of infrastructure for our clients.

Oct 2018 – Nov 2018 · 1 mo · United Kingdom · Remote

• This was my first experience with a Fintech startup.
• At KyoLAB, I helped in DevOps by deploying a containerized version of our collaboration tools on our own dedicated servers. Containers were clubbed and managed using docker-compose.
• I was also responsible for creating a continuous integration and delivery pipeline for Android applications using a scalable and containerized version of Jenkins.
• Thus, most of the time here I found myself playing with Linux.
• Besides the occasional DevOps, I did JavaScript and React.js to help build a SPA too!

Oct 2017 – Mar 2018 · 5 mos · On-site

• I lead the entire team of GraphicWeave in this brief exciting period.

Apr 2017 – Sep 2018 · 1 yr 5 mos · On-site

• My roles at GraphicWeave included designing cloud native systems that were vertically as well as horizontally scalable. I developed, designed and bridged complete solutions on different platforms from the ground up. I worked on code monoliths to distributed systems employing micro services and continuous integration.
• I coded for the web and made dynamic UIs using React.js & Vue.js.
• I did mobile development on the side too. Here, I was also a native as well as a hybrid mobile developer.

Apr 2016 – Mar 2017 · 11 mos · On-site

• This was my first job and back then when it was a startup.
• I was hired as a back-end programmer back then to build projects and help with the development of web and mobile applications. I mostly looked after server-side code that was optimized (not over optimized though, haha!) and modular, which followed the principles of high cohesion and low coupling. I was part of a team that had years of professional experience.

Oct 2017 – Feb 2018 · 4 mos · Chandigarh, India · Remote

• Meego is a helpline service that connects tourists and travelers to government registered tour guides that are nearby. It also offers points of interest service that connects users to popular places that are nearby.
• I designed and engineered this project myself. During my 5 months tenure with Meego, I handled all of its API development, almost all of its app development and DevOps.
• Meego has two apps, one for the user and one for the tour guide.
• One of the apps is cross-platform and works on both Android and iOS. It is written in JavaScript using React Native and Expo.
• The other app works only on Android and is a native app written in Kotlin.
• APIs were written in Node.js with Hapi.js as the framework of choice and MongoDB as the backend database. All of the backend and web apps are deployed inside scalable docker containers with API documentation done using Swagger.
• Other technologies involved:
• Remote logging and crash reporting - Sentry and Crashlytics.
• Continuous deployment (CD) - Docker Cloud.
• Continuous deployment of apps - Fastlane.
• Analytics - Segment and Mixpanel.
• State management - Redux and SnappyDB.
• API - RESTful.
• Real-time communication bus - Deepstream.
• Payment gateway - Razorpay.