Umesh Dangat
Principal Engineer at Yelp
San Francisco, California, United States
Summary
🤩
Rockstar🎓
Top SchoolUmesh Dangat is a Principal Engineer based in San Francisco with nine years of focused experience building search infrastructure and scalable backend systems, currently leading engineering efforts at Yelp. He has a strong foundation in large-scale product engineering from roles at Quid, Symantec, Teradata, and Infosys, and holds a Master's in Computer Science from USC. Umesh contributes to open-source search tooling—most notably enhancing the elasticsearch-learning-to-rank plugin to enable query-time feature selection, query-time variables, and script parameter indirection—improving relevance customization for production search. Known for bridging research ideas and production constraints, he combines deep systems expertise with practical improvements that make ML-for-relevance more flexible and reusable in real-world search applications.
9 years of coding experience
12 years of employment as a software developer
Masters Computer Science, Masters Computer Science at University of Southern CaliforniaBachelors Computer Engineering, Bachelors Computer Engineering at Savitribai Phule Pune University
Github Skills (10)
plugin-development10
javas10
elasticsearch10
elasticsearchquery10
feature-engineering10
aws-elasticsearch10
java10
elasticsearch-api10
amazon-elasticsearch10
api-design8
Programming languages (1)
Github contributions (5)
o19s/elasticsearch-learning-to-rank
May 2018 - Aug 2019
Plugin to integrate Learning to Rank (aka machine learning for better relevance) with Elasticsearch
Role in this project:
Back-end Developer
Back-end DeveloperContributions:16 commits, 11 PRs, 4 pushes in 1 year 3 months
Contributions summary:Umesh primarily focused on enhancing the `elasticsearch-learning-to-rank` plugin's core functionality. They implemented the ability to select a subset of features at query time, introducing a new parameter to the `sltr` query. Additionally, the user added support for query-time variables within the `derived_expression` feature and implemented indirection of names via `extra_script_params` in ScriptFeatures, facilitating reuse of existing scripts. These changes improved the flexibility and customizability of the plugin's relevance scoring capabilities.
machine-learningelasticsearch-pluginelasticsearch-pluginslearning-to-rankrelevance
duckbills/platypus
Feb 2020 - Oct 2022
A high performance gRPC server on top of Apache Lucene
Contributions:77 pushes, 34 branches in 2 years 8 months
protobufgrpcgrpc-serverapacheon-server
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.