Aug 2020 – Dec 2025 · 5 yrs 4 mos · Sunnyvale, California, United States · On-site

• I proposed AI co-workers to Satya Nadella and inspired a Microsoft wide re-prioritization around agentic AI. Drove multi Billion $ ideas (CVP/Technical Fellow level ideas) such as disrupting Azure using GPU cloud marketplace leveraging hybrid cloud infrastructure and digital AI coworkers using agentic AI by gathering volunteer teams. My reviews from executive leadership said, "Microsoft continues to thrive because of people like you", "This is really smart project". I was the 1st prize winner from Twenty Thousand teams and Seventy thousand hackers in 2024. 1st prize hackathon winner for Azure organization for 2023, 2024. Defining the future of work using LLM/Agentic Computing/AI teammates. Worked on Developer productivity, video ai, edge ai, security ai in the past.
• I have worked on solutions at the convergence of LLM Code generation, Data Pipelines, Hybrid Infrastructure, Copilot tech, Agentic AI systems, OS, Autonomous Systems, Deep Reinforcement Learning, Security, Hybrid Cloud, Kubernetes, Confidential Computing, Responsible AI, Metaverse, , LLM, Edge Computing, Autonomous Edge, Graph Search, Mixed Reality, ML Ops, Sensor Fusion, Active Transfer Learning, 3D Reconstruction, GAN, Data Analytics, Model Compression, Sovereign Cloud, Air gapped Cloud, Hybrid AI, Generative AI, Analytics, and Video AI.
• Some projects:
• a) Primary author of 3 patents filed with USPTO and 9 patent proposals in areas such as Gaming, Data Poisoning, Generative AI, Adaptive Operating Systems, Multi-Modal Data Compression using Deep Learning (In 2020), novel Self-Driving tech etc.
• c) Developed prototypes and products in Generative AI in 2021, Confidential Computing in 2023, award winning idea ahead of the ecosystem.
• d) Classified projects in Video AI, LLM's, Agentic Computing, Edge Computing, Autonomous Systems, and Hybrid Computing.
• e) Presented thoughts to multiple CVP's, EVP's, Satya, Scott G, and Jay Parikh.

Aug 2018 – Apr 2020 · 1 yr 8 mos · San Francisco Bay Area

• At Macy's, I have solved interesting Data and Machine Learning problems using techniques from Vision, NLP, Cloud Infrastructure, Knowledge Graphs, Query Understanding, Search, Personalization, Deep Learning, Big Data, and Deep Reinforcement Learning working with several leaders with 100+ engineers across the organization.
• Some solutions, I worked on are:
• a) Automated generation of outfits for fashion using Vision and NLP techniques.
• b) Generating query suggestions and item recommendations using Graph and Search Techniques.
• c) Improve recommendations using Deep Learning.
• d) Explored Pricing Adjustments using Deep Reinforcement Learning.
• e) Evangelized unified DL Embedding techniques for different use cases.
• f) Personalized Reranker for eCommerce Recommendations using DL.
• g) Enterprise-wide Machine Learning platform on Google cloud for different ML use cases.

Dec 2016 – Jul 2020 · 3 yrs 7 mos · San Francisco Bay Area

• I have deep domain and product experience in Search, Ads, Recruiting, Marketplaces, eCommerce, and Social Networks, and technology expertise in Artificial Intelligence, Chatbots, Vision, NLU, Multi-Modal Dialog Engine, Search, Deep Learning, Big Data, Cloud, Algorithms, and Scalability. Please let me know if you need help with your organization. Please see some of my work, as follows:
• Contiq (October 2015 - January 2016): As a consultant CTO, I created an architecture plan for Content Search and Recommendations for automating the generation of sales presentations.
• Wells Fargo (December 2016 - July 2018): I took a consulting job at Wells Fargo as part of my startup journey at Voicy.AI.
• In my consulting role, I designed and architected an omnichannel AI platform for multiple use cases at Wells Fargo. Some of my responsibilities are:
• a) Entity Recognition Engine building up on Dictionary, Deep Learning, and Rule-Based logic.
• b) Intent Recognition Engine leveraging Search and Deep Learning Techniques.
• c) Supported features such as Context Switching, Mis-Spellings, Multiple Intents, and Memory to the Virtual Agent.
• d) Overall AI Architecture for Wells Fargo.

Oct 2014 – Sep 2015 · 11 mos · San Francisco Bay Area

• I built and led a team of engineers to solve search ranking problems for Amazon's Digital properties including Amazon Video, Kindle, and Music. My team has pushed new Machine Learning models for Search Ranking of Video and Music content. I also helped to release Amazon Digital subscriptions in Europe and Asia.
• I personally wrote 3 patents focusing on implicit profile personalization for Video streaming sessions, the applicability of Holograms in an eCommerce context, and usage of Machine Learning techniques to identify and tag videos to generate automated personalized video trailers.

Mar 2014 – Sep 2014 · 6 mos

• Helped with Search Platform re-architecture for local. Developed code base for feature engineering, buy predictions, data pipelines etc.

Sep 2012 – Apr 2014 · 1 yr 7 mos · Sunnyvale

• My responsibilities included helping with algorithms to optimize revenue and relevance metrics for search queries, coming up with frameworks to increase monetization on partner sites, and managing Search and Ads auction fleet of servers. Some of my/team's accomplishments were:
• a) Migration of search platform from Lucene to Solr
• b) An increase of 10% job view per unique visitor with a series of relevancy changes in the search platform.
• c) An increase of 12% revenue per unique visitor by doing query expansion and increasing the recall size to the second pass algorithm.
• d) Pushed optimizations such as aggressive caching and continuous indexing.
• e) Evangelized pub sub architecture across the company.
• f) Led development of Experimentation Framework for search.
• g) Sourced and hired a great addition to the team.

Jun 2010 – Dec 2011 · 1 yr 6 mos · San Jose, CA · On-site

• Came up with a way to aggregate activity across internet of a user and use it to generate recommendations/gifts. Pioneered social commerce before the behemoths.

Jan 2010 – Jan 2016 · 6 yrs

• Groupon (April 2014 - September 2014): Helped migrating to a newer search infrastructure and developing a framework to improve Search Relevance.
• Jobsmunk (May 2014 - December 2014): Agentic AI recruiter in 2014. Jobsmunk helps to connect job seekers and employers using resume and job descriptions. Built a career graph from unstructured content in the resume.
• FirstRadius (March 2010 - December 2014): FirstRadius.com uses social profiles to target content. We pioneered and patented techniques for Custom Audience, Ad Targeting using id based behavior aggregation, Social Gifting ideas in August 2010 (Compared to Facebook's Custom Audience patent from June 2011), Social Search (January 2011), etc as part of the journey.
• We developed TopGifter.com, an early gifting recommendation engine which shows relevant gifts aggregated from 1000's of Retail Stores, your friend likes using his/her social profile to validate our ideas in social recommendations.
• Maansu (September 2010 - March 2012): I helped developing an Amazon-like business model with a few engineers and sold millions of products to thousands of customers in India. We pioneered Dynamic Product grouping to optimize shopping costs and Search Assistance before Flipkart and Snapdeal.

Apr 2009 – Sep 2012 · 3 yrs 5 mos

• I helped pushing personalized content across ebay. Some of the important projects, I executed with the help of a wonderful team are:
• a) Merchandizing recommendations across eBay.
• b) Integration of merchandising recommendations on acquisitions such as Stubhub.
• c) Email Recommendations platform.
• d) Ads Personalization engine.
• e) Merchandizing recommendations on MyeBay.
• f) Dynamic and scalable platform for serving content recommendations for Billion hits per day.
• g) Merchandizing recommendations in Emails.
• h) Prototype for shopping comparison on Pinterest data, leveraging Lucene, Solr and Nutch.

Sep 2008 – Apr 2009 · 7 mos · San Mateo

• I worked on Rule Engine for the Data Loss Prevention group. The rule engine helped organizations to protect sensitive data.

Apr 2008 – Sep 2008 · 5 mos · San Francisco Bay Area

• I worked on Evault's Search and Information Discovery product.

Apr 2004 – Apr 2008 · 4 yrs

• I joined Yahoo! Music after the acquisition of Launch group by Yahoo! I was lucky to be mentored by some of the greatest technologists at an early age. Some important products, I worked on are:
• a) Music Content Integration: I ingested millions of records from multiple partners to build an integrated view of the best music metadata on the web. Some of the problems I solved were Fuzzy Search and Duplicate Resolution. I also wrote an implementation to do efficient pre-processing of records for searching in SQL with complex query optimizations.
• b) Content Publishing: I leveraged publish-subscribe messaging infrastructure of Yahoo music and scaled the content publishing process to millions of users.
• c) Music Search: Developed a vertical search engine for Yahoo Music. I innovated a solution for the problem of misspellings in name queries. This approach was co-incidentally adopted by Solr after four years.
• d) Music Personalization: Worked on solutions to personalize Music and Video for Yahoo! users.
• e) Yahoo Ads: Denormalization framework for Yahoo! Ads. I was lucky to be one of the early users of Hadoop and Map Reduce abstraction. I used the Map-Reduce paradigm to join records from multiple databases to create an Inverted Index. This index was then used to serve ads to users using Information Retrieval techniques.
• f) Yahoo! Display Ads: I helped writing an application to cross-sell display ads from Newspapers consortium on Yahoo!
• I also patented the ideas of Search Re-Targeting (Patented in 2007, Idea gained traction in 2010-2011), Ad Measurement across the web (Patented in 2008, gained traction in 2012, mass adoption by Google in 2017) for Yahoo!

Sep 2003 – Mar 2004 · 6 mos · Bethesda, Maryland

• I made enhancements to the testing framework for hotel reservation component of Marriot.com.

Oct 2001 – Jul 2003 · 1 yr 9 mos · Tucson, Arizona Area

• I conceptualized and developed an implementation of Ajax (The Ajax term was coined after several years of my use of the technology) application using Microsoft technologies.
• I used my self-taught knowledge in server-side programming, client-side programming, Java CICS framework, Fuzzy Search, and Query Tuning to architect and write a web application to transfer degrees from other colleges to the University of Arizona.