As a Backend Staff engineer on Uber’s API product for Delivery called Uber Direct, I identified many improvements, brought buy-in from stakeholders and got these projects funded to achieve organizationally important goals. I work with several managers, engineers, Product Managers in Uber Direct on its architecture and its future direction. Other than regularly involved in code/design reviews, providing guidance to junior/senior engineers and project scoping for a 10 people engineering team, I have impacted the entire product through my ideated projects:* Built relationships with Maps org at Uber and influenced their roadmap in order to deliver on my technical vision for our product. * Identified a Maps feature that was built for customer-facing UI and influenced our product managers to retrofit that to our API product; this improved our delivery completion rate.* Ideated, designed and built a generic request/response Hive Logging framework used by several members of the org to run analytics on request/response data.* Identified various issues affecting the Availability/Uptime for our APIs. Got many of the org’s 50+ engineers involved in fixing these issues to meet our 99.99% SLOs.* Identified a way to increase Direct’s orders by using Isochrones to capture demand that were rejected due to distance based restrictions for our Merchants.* After the acquisition of Postmates, I worked on Postmates integration with Uber. I worked with the UberEats Membership team to migrate Postmates Unlimited subscriptions to Uber's Eats Pass. Among solving other challenges, there were challenges in finding a solution for migrating Apple/Google pay subscribers in Postmates and I helped come up with a design for the same. Apple/Google pay subscribers made up a significant portion of Postmates subscribers.

Led the effort to increase the number of restaurants available in the customers feed while optimizing for driver travel time, using isochrones. * Collaborated with various stakeholders (Finance, DS and ENG) to define/implement experiments to pick the best pricing strategy with isochrones in various markets to maximize GMV.* There were performance bottlenecks for isochrone queries in Postmates’ Elasticsearch instance which I helped improve using various optimizations which helped unblock the project.

* Led the effort to design, implement and launch Netskope’s Auto-remediation product for AWS using python. Worked with the Product managers and UX team to design the v1 of this product. This allows customers to correct issues in their AWS instances. For example, this product allows automatically rectifying a publicly exposed/accessed S3 bucket by setting its access level to private automatically.

* Owner of the queryengine used for powering Netskope's reports and dashboards for showing analytics data. The reports and dashboards support querying on large datasets from two different NoSQL datastores – MongoDb and Druid. In addition to this, it supported several in-memory aggregations and filters.* Led the effort to improve the performance of backend queries in the Queryengine for our new UI using a new datastore named Druid.* Contributed ideas to the UX team to redesign Netskope’s UI while keeping the UI performant in terms of backend queries - https://www.netskope.com/blog/enabling-design-driven-real-time-insights-at-scale/

Designed and implemented an Observability framework for reporting issues in real-time on Netskope’s appliances. Before this was developed, the customer support team had no visibility into the issues on appliances until the customer reported the same.

* Worked on various projects using Open source software like MongoDb, Kafka and Redis. Worked on a data ingestion service, a service for calculating hourly and daily summaries and a distributed query engine to run reports and ad-hoc queries on the raw/summarized data.* Re-architected and scaled a service that has functionality of https://haveibeenpwned.com/but only for enterprise customers. This allowed enterprise customers to get notified when their employees’ emails were found in a breach. The service was able to analyze hundreds of millions of events every day. Kafka, Mongodb and python were used for this.

As part of the core-infrastructure team of Amazon Photos, contributed to the redesign and rewrite of the platform backend using Java and AWS. Built a data-pipeline using AWS Kinesis, SQS and CloudSearch to keep CloudDrive's DynamoDb and CloudSearch stores in sync in real-time. A presentation on the architecture is here: https://www.youtube.com/watch?v=R2pKtmhyNoA

Worked as part of the Tax team to build software for calculating taxes for Amazon checkouts.

Wrote extensions to EDG's C/C++ compiler frontend so that Coverity's static analysis for bug detection can be performed on programs written for a variety of C/C++ compilers.

HPC and Compiler projects.

Implemented Automatic graph layout for LabVIEW - a graphical programming language. This project was nominated for the “Best Innovation of 2008” award at NI, one of 5 such projects across NI worldwide. This also resulted in a patent.Please see a demo here - https://www.youtube.com/watch?v=SasNXh_fk8E

Developed a DataGrid that could host controls in .NET Compact Framework.

.NET API for NI-Digitizers