Jul 2022 – May 2024 · 1 yr 10 mos · Karnataka, India

• Improved the payment method recommendations on the Payments Selection Page (PSP) by building a backend API that recommends payment methods based on multiple strategies like frequency, recency, etc, within defined SLA constraints. It involved working in a distributed architecture ensuring seamless scalability and performance. Moreover, I designed a flexible service interaction model, capable of accommodating diverse business use cases, thus laying a foundation for future extensibility and adaptability.
• Revamped the user experience of the Payments Selection Page (PSP) through a series of strategic UI enhancements. This initiative entailed refining the messaging across multiple payment methods to enhance clarity and user understanding. Additionally, I streamlined the rendering process for various payment methods by integrating existing logic into a unified template, ensuring consistent and cohesive presentation throughout. These improvements not only elevate the visual appeal of the PSP but also enhance usability and coherence for users navigating different payment options.
• Improved the availability and resiliency of various services that handle millions of transactions per day. This involved developing scaling formulas, performing load tests, and identifying and working on tasks that would enhance the availability of the systems. This proactive approach not only minimized downtime but also optimized system performance, ensuring a robust and reliable user experience even during peak usage periods.

Jan 2022 – Jun 2022 · 5 mos · Karnataka, India

• Built a standalone automation tool from scratch that attaches the relevant logs, metrics, and other critical information to the incoming production issue within 2 minutes with zero manual intervention.
• SNS listens for any incoming production issue events and publishes this to SQS. A Python script performs long polling on SQS to get messages. From this message, it extracts the SNS event and if that event is for a new ticket, further processing is done. The ticket is parsed to get the alarm name that triggered the ticket and the timestamp of the alarm creation. Using the alarm name and timestamp, details of the monitor are fetched. Based on these details, the query is performed on Cloudwatch to get the required logs. Since the log size is large, a link is generated for these logs. Finally, the ticket is updated with the data that can help the team member find the root cause of the issue.

Jun 2021 – Jul 2021 · 1 mo · Chennai, Tamil Nadu, India · Remote

Mar 2020 – Apr 2021 · 1 yr 1 mo · Vellore Area, India

Jan 2019 – Oct 2020 · 1 yr 9 mos · Vellore Area, India