Jul 2021 – Jul 2024 · 3 yrs · Bengaluru, Karnataka, India

• Worked on Cloud Data Fusion, a GCP product which allows users to build data pipelines using a visual interface with drag-and-drop operations.
• Modified the internals of Data Fusion to enable optimizations in aggregation operations.
• Collaborated with several teams to meet Google-wide horizontal compliance requirements issued by the US Federal Government.
• Projects:
• Implemented Access Transparency for Cloud Data Fusion, enabling customers to see when Googlers have accessed their Data Fusion instance in their logs. This included mapping the resource identifiers for dependent resources used internally by Data Fusion (which are not customer-visible) to the resource identifier of the customer's Data Fusion instance.
• Added SQL capability to the Wrangler plugin's Precondition filter in addition to the existing JEXL-based interface.
• Implemented the Transformation Pushdown feature which allowed an operation in Data Fusion to be pushed down to another execution engine (such as BigQuery) for the Group-by, Deduplicate and Wrangler plugins. See the documentation here: https://docs.cloud.google.com/data-fusion/docs/concepts/transformation-pushdown

Apr 2019 – Jul 2021 · 2 yrs 3 mos · Bengaluru Area, India

• Developed big data applications for Supply Quality for Advertising, which involves detecting whether bids, clicks and impressions for advertisements are human or robotic at a scale of about 1.3 million average qps with a 40 ms latency requirement.
• Projects:
• Away teamed at the Amazon Ad Exchange team to add logic to filter incoming bid requests for ads on the basis of app-ads.txt files (for mobile traffic) in addition to the existing logic for ads.txt files (for web traffic). See the Amazon launch announcement here: https://advertising.amazon.com/en-ca/resources/whats-new/app-ads-txt-support-in-amazon-dsp
• Created a fetcher for fetching the ads.txt files for desktop websites for which bid requests are received and mentored another SDE to create a similar fetcher for app-ads.txt for mobile apps.
• Enabled pseudo-anonymization of PII data in the real-time persistence pipeline for incoming clicks and impressions data on the Amazon.com website.

Aug 2015 – Apr 2019 · 3 yrs 8 mos · Bengaluru Area, India

• Performed Data-as-a-Service development for the Visa Data Manager application.
• Reimplemented existing Ab Initio graphs performing ETL processing in Hadoop using Hive for the restructuring of Visa's tech stack under the Long-Term Architecture project.
• Projects implemented:
• Designed and Implemented an Event Detail report for the VDCA mobile application which enriched missing data in user generated events (logins, payments, etc). Used Hive
• Migrated the Global Deduplication of Transactions (GDT) application from Ab Initio to Hadoop. This involved finding duplicates in transaction data and sending these duplicates to banks. Used Hive.
• Changed the data source for the VDCA application from Oracle to Hive for over 100 tables without
• affecting the underlying process reading from those tables. Used Hive and Bash
• Built a generic process to extract over 120 ‘code tables’ in the form of excel files on a daily basis and
• load the contents in Hive and DB2 for further ETL consumption. Used Java, Bash and Hive
• Converted an existing Ab-Initio process that aggregates transaction data by cardholder to Hadoop by
• reimplementing all business logic. Used Hive
• Created a generic utility to transfer files across servers with built in automatic file-level retries on
• failure. Used Bash

Jul 2013 – Jun 2015 · 1 yr 11 mos · Mumbai Area, India

• July to Dec 2013 under Prof. D. B. Phatak
• Investigated multilingual text editors for Prof. Phatak’s Teach 10000 Teachers Programme.
• Jan to May 2014 under Prof. Pushpak Bhattacharyya
• Performed lab evaluation and paper checking duties for a 100 student course on Artificial Intelligence.
• Jul to Dec 2014 under Prof. Pushpak Bhattacharyya
• Performed assignment evaluation and paper checking duties for a 140 student course on Natural Language Processing. Included a guest talk on 'Classifiers for Sentiment Analysis'.
• Jan 2015 onwards under Prof. Jayendran Venkateswaran
• Performed paper checking duties for a 170 student course on Probabilistic Models.