Adwait Tumbde
Engineering Manager at Meta
Sunnyvale, California, United States
Summary
🤩
Rockstar🎓
Top SchoolAdwait Tumbde is an engineering manager with 15+ years building and scaling distributed systems, ML infrastructure, and data platforms, currently leading a multi-team org at Meta that powers hundreds of Ads ranking and recommendation models. He has a strong record of delivering 0→1 platforms—most notably founding HawkEye, an ML observability system that cut debugging time by 80% and drove hundreds of millions in annual savings—and driving platform changes that produced an estimated 35–40% incremental revenue impact for Ads. Previously at LinkedIn he scaled a data access and virtualization platform (Coral) to company standard, improving query portability and reducing latency by up to 100%, and contributed to Apache Pinot and large-scale OLAP advances. Comfortable at the intersection of product, research, and engineering, he excels at recruiting and growing senior technical leaders and pushing architecture-level optimizations across compute, storage, and data access layers. Based in Sunnyvale, he combines deep distributed-systems craftsmanship with hands-on open-source contributions to query translation and real-time analytics.
10 years of coding experience
13 years of employment as a software developer
MS, Computer Science, MS, Computer Science at University of Wisconsin-MadisonBS, Computer Science, BS, Computer Science at National Institute of Technology, TiruchirappalliSomalwar High School
Github Skills (12)
apache-calcite10
data-storage10
javas10
apache-pinot10
sql10
back-end-development10
java10
calc10
multithreading9
performance-optimization9
memory-management9
unit-testing9
Programming languages (3)
Github contributions (5)
linkedin/coral
Jul 2017 - May 2019
Coral is a translation, analysis, and query rewrite engine for SQL and other relational languages.
Role in this project:
Back-end Developer
Back-end DeveloperContributions:91 commits, 2 comments in 1 year 10 months
Contributions summary:Adwait primarily contributed to the development of the Presto dialect within the Coral translation engine. Their work included creating the `RelToPrestoConverter` and `PrestoSqlDialect` classes, and adding unit tests to validate the conversion of Calcite RelNodes to Presto SQL. They focused on implementing support for various SQL clauses, including subqueries and joins, which involved modifying and extending the core functionality of the Coral project to generate Presto-compatible SQL. This work aimed to enhance Coral's capabilities for query rewriting and translation across different SQL dialects.
queryrelationalcoralflink-sqltranslation
apache/pinot
Sep 2015 - Sep 2017
Apache Pinot - A realtime distributed OLAP datastore
Role in this project:Back-end Developer
Back-end DeveloperContributions:139 commits, 160 PRs, 243 pushes in 1 year 11 months
Contributions summary:Adwait primarily worked on improving the Apache Pinot datastore's backend. They fixed thread-safety issues in the MmapUtils class by switching to a synchronized map. The user also made the TimeRetentionStrategy constructors consistent by fixing potential null pointer exceptions for bad time unit values. Additionally, they addressed memory leaks in the RealtimeSegmentImpl class and incorporated performance optimizations for aggregation functions. The user also contributed to updating the table format to v3.
realtimedata-streamolapapachedatastore
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.