Feb 2025 – Jun 2025 · 4 mos

• Training infrastructure

Nov 2022 – Feb 2025 · 2 yrs 3 mos

• Leading a team of talented ML Engineers to build world class ML & Data infrastructure that helps build, optimize and deploy self driving models of Nuro fleet.

Mar 2020 – Oct 2022 · 2 yrs 7 mos

• As a Principal Applied Scientist, lead AI Engineering for smart features in Microsoft Family Safety products.
• Inception, prioritization & execution of AI features.
• Mentoring early career folks, helping them with career guidance and upskilling in Engineering and AI.
• Created AI Playbook for streamlining AI project development and Infusing data/ML thinking to partner teams in product, engineering & design.
• Guiding team with technical directions for Data/AI as an architect.
• Integrated various content safety models in Edge to protect kids from harmful content (images, search queries, text) while browsing.
• Operationalized short text harmful content classification service serving 20+ million requests/day.

Feb 2018 – Mar 2020 · 2 yrs 1 mo

• Responsible for all data and AI features for Microsoft Money and MileIQ products.
• Lead for all the smart features end-to-end, from ideation to production & post-production. Some examples of our consumer financial insights include - large purchases, subscription detection, transaction categorization, merchant name clean up.
• As a privacy lead, built tools and processes to ensure compliance for financial data while maintaining accessibility for AI and data science use cases.
• Before that, as part of MileIQ team
• Drive agile mindset, processes and automation for ML/data features.
• Built AI & analytics platform (TensorFlow, Spark, Jupyter + Kubernetes + Operator framework) to streamline various use cases with standard workflows across various data teams. There are multiple internal teams leveraging the platform to speed up their AI workflows.
• Built offline experimentation engine to allow rapid iterations on drive algorithm.
• Architect for all things data

Apr 2016 – Feb 2018 · 1 yr 10 mos

• As part of Azure HDInsight team,
• Led query engine performance & benchmarking
• Improved Spark optimizer
• Presto integration with HDInsight
• Released Interactive query (LLAP) for hyperscale analytics
• Multi-tenant Hive job service on Azure Data Lake Analytics platform
• Enterprise Security Package for Azure HDInsight
• Helping some of the largest Azure HDInsight customers with their large scale Analytics initiatives

Apr 2014 – Apr 2016 · 2 yrs

• Research in Data and Distributed Systems

Nov 2013 – Apr 2014 · 5 mos

• I worked with Dr. Kaushik Rajan to build ways of constructing critical path models for Yarn, Hive and Hadoop jobs using fine grained statistics (codename PerfOrator project). We used it to predict execution time and resource requirements of jobs and integrate it with reservation based scheduler in YARN.