SAMMARTH KAPSE

Automatic Reconciliation Tool (ART) is responsible to ensure the money flow in Razorpay. Automated reconciliation process for 30+ gateways and channels in Razorpay’s RX and Payment Gateway operations, accommodating unique configurations for each. Reconciling bank statements against Razorpay’s records, carefully managing payment edge cases, and diligently identifying potential money leakages. With volume of over 70 lakh transactions daily. The system’s APIs facilitate gateway configurations, alerting mechanisms, result sharing, and the seamless onboarding of new gateways along with their dependencies. Implemented distributed computing using Spark on Databricks clusters for efficient large-scale record processing. Tools/ Technology: Backend Framework: Flask, PHP; Databases: MySQL, PostgreSQL; Big Data: PySpark, Databricks; AWS Lambda, S3, SQS, Kafka, CronJob.

Developed a project aimed at enhancing development productivity by analyzing the Pull Request (PR) cycle, both historically and in real-time. Extracted historical PR data from Github logs and refined it to create insightful metrics and visualizations. Implemented real-time monitoring of all Pull Requests using Github Webhook. Assessed PR quality based on multiple parameters such as changes made, comments and closure duration. Tools/ Technology: Github APIs, Github Webhook, AWS Lambda, GoLang, and Python.

Developed a custom React rendering engine that generates dynamic forms from JSON input files. Allowing users to specify required form components in JSON files, resulting in dynamic form rendering through React. Transformed given components into custom React components. Used Elastic Search as the storage system to store JSON files and user data. Tools/ Technology: React, Java, Elastic Search.