Jan 2017 – Jan 2024 · 7 yrs · San Francisco Bay Area

• Key Achievements -------
• Leadership
• Built & led high-performing ML teams in India & US, delivering strategic initiatives with significant business impact.
• Established the first ML team in India, setting up recruitment, interview processes, and hiring top talent.
• Mentored over 2 dozen engineers, fostering their growth and career development.
• Technical Expertise
• Developed planet-scale ML models for diverse applications, including next activity prediction, map error detection, support automation, workforce forecasting, and high-value customer identification.
• Pioneered NLP & Generative AI at Uber, driving projects that improved customer satisfaction by 8%, reduced triage time by 90%, and optimized support workforce by 25%.
• Initiated and delivered multiple "0 to 1" projects with high business impact, including Project Next (patent-pending) for next activity prediction and Project Wallaby for optimized delivery times.
• Communication & Collaboration
• Effectively communicated technical concepts to stakeholders and got leadership buy-in for multiple 0 to 1 efforts and vision.
• Presented work to the CEO/leadership and multiple company-wide talks.
• Led AI charter for one org and a key contributor to the NLP infra roadmap for the entire company, rallying multiple teams together.
• Single-handedly created and conducted an ML course attended by 150+ engineers and PMs.
• Most recently: TL for GenAI efforts -------
• Responsible for transforming the customer support org using AI. Before I joined, it was mainly an engineering org without a clear ML strategy. Came up with the ML roadmap which is now being followed even four years into it.
• Pioneered LLM adoption, launching impactful projects and expanding Uber's technical capabilities.
• Led Generative AI efforts, delivering 3 successful projects and shaping the team's strategy.
• Actively presented cutting-edge work at company-wide and leadership meetings.

Jan 2014 – Sep 2014 · 8 mos · San Francisco Bay Area

• Product listing ads (PLAs) are the sponsored product ads for queries like "canon eos". Along with a team of 40+ engineers, I was responsible for PLA and shopping.google.com search quality. This involved everything from understanding the user query to ad retrieval, prediction models, auction and triggering.
• My goal was to provide the best product shopping search experience in the world for our users, create value for merchants and generate revenue for Google. I also focused on our international growth strategy for tablet and mobile working closely with engineering and local sales.
• I was on this role for a within Google rotation program and during the time I was there, my key achievements were:
• Came up with projects that led to significant (double digit) growth in PLAs on mobile and tablet.
• Came up with a roadmap for our strategy in APAC where we had some issues.
• Came with the proposal for Google shopping's next generation query and product understanding system - a system which would deeply understand what a user wants from a query and what a product offers. This wasn't finished while I was there, but we were able to form a new team and start working on it.
• Help eng team set milestones and execute on them.
• Eventually, decided to come back to engineering.

Feb 2012 – Jan 2017 · 4 yrs 11 mos · San Francisco Bay Area

• I work on Google's web search ranking algorithm. Our goal is to improve the quality of results we show to our users for any given query. This involves building scalable systems that can understand the web and the user intent and provide best results across geography and languages.
• Main skills involved are web scale data mining, big data analysis, feature engineering, machine learning and building fast parallel multithreaded distributed system.
• My work has had a major impact over the entire search stack from document processing at the crawl time to their ranking at the serving time.

Feb 2012 – Jan 2014 · 1 yr 11 mos · San Francisco Bay Area

• I work in the web search ranking team, where we try to understand the web and help our users get to what they need.

Aug 2010 – Dec 2011 · 1 yr 4 mos · Mumbai

• VocabMax was an e-learning tool to help students / professionals improve their english vocabulary and help them prepare for exams like GRE / CAT / GMAT.

Jul 2010 – Feb 2012 · 1 yr 7 mos · Mumbai Area, India

• Worked as a Quant for the Strats and Modelling Division.
• Skills involved: Derivative pricing, Market Simulation and risk analysis, Interest rate modelling using monte carlo methods, unified product description language designing, Stochastic calculus.

May 2009 – Jul 2009 · 2 mos · Vancouver, Canada Area

• Devised a new framework for recommender systems that is much more exploration friendly than the existing recommender systems like Yahoo! Movies, Amazon etc.
• Presented the work at MITACS Globalink Research Symposium 2009 held at University of British Columbia, Vancouver, Canada

May 2008 – Jul 2008 · 2 mos · Paris Area, France

• Designed and implemented algorithms for efficient data storage and retrieval in a peer to peer network.
• Project was among the 30 selected (out of 109 proposals submitted worldwide) for demonstration in Very Large Data Bases (VLDB) Conference 2008, held in Auckland, New Zealand