At Intuit, my team and I are pioneering real-time identity resolution across an expansive ecosystem, leveraging advanced technologies like Apache Flink and Kubernetes. Our work empowers critical use-cases for products such as TurboTax and QuickBooks.
Building real-time identity resolution, using Apache Flink, Kafka, GraphQL, Springboot, PostgreSQL, Entity Resolution sytem to enable use-cases across the Intuit ecosystem, comprising Turbotax, Credit karma, Quickbooks & Mailchimp.
Lead the development of a low latency & resilient GraphQL micro-service using Neflix DGS framework. Deployed and scaled the service on Kubernetes.
Designed and implemented various features to support Authorization, API maturity and various use-cases, along with making the service production ready through performance testing, tuning, FMEA testing and setting up monitoring dashboards and alerts.
Designed and implemented an algorithm in Flink using Java to resolve results from identity resolution system to generate a new identity, called Resolved Identity, that can represent a same person or company with different accounts throughout the Intuit ecosystem, enabling seamless login, marketing, and personalization use cases.
Designed and implemented a persistence library using Java persistence API, which allowed Data model definition, DDL generation and provided persistence APIs, that are used across all the services and pipelines of the Id-resolution platform.
Designed and developed a new notification system in Flink to publish Resolved Identity changes to consumer teams within Intuit, enabling a new access pattern that can send updates in real-time via Kafka.
Worked as a part of the Data-bridge team, enabling real-time data flow between Intuit and its subsidiaries, Credit Karma and Mailchimp, in a secure and managed way using Apache Flink, Kafka, and Google Pub/Sub.
Improved the resiliency of Apache Flink applications by developing a feature to use AWS S3 as a backup to retrieve Data-bridge configuration files in case of Intuit config service being unreachable.
Developed a Jenkins pipeline to automate end-to-end Flink job deployments and rollbacks along with relevant configuration updates, reducing manual time required by 70% from 10 minutes per job to 3 minutes in total.
Set up AWS Lambda functions to refresh Intuit private keys and pull new versions of public keys from Credit Karma, enabling automatic key rotation and eliminating downtime due to manual key rotation.
Worked as a part of C360 team, responsible for creating a unified profile of users by combining data from high volume data sources, and providing a reliable data platform to power business intelligence and analytics at Intuit.
Enhanced the existing Stream processing platform to process new attributes in data using Apache Flink and store them into DynamoDB, reducing the time to execute data-based experiments from 2 weeks to 2 days.
Engineered a project to derive features and provide real-time analytics from time-series data using Apache Druid, aimed at providing better personalization to users by capturing their journey over time.
Designed and launched a monitoring system using AWS Cloudwatch, Lambda, and Python, which reduced the time to detect and investigate EMR errors from multiple hours to less than 30 mins.