Aug 2020 – Jun 2021 · 10 mos

• Part of Google Maps team, with focus on the Vehicle Map Service (VMS). VMS is an in-car data exchange service that shares road and navigation data with the vehicle’s real-time operating system, allowing many vehicle components and systems to behave more intelligently as they gain awareness of the road around them. As part of the VMS team, I was leading the VMS tooling efforts involving building visualization and analysis tools to easily simulate drives, understand the data being published to the vehicle side by side with signals from the vehicle to gain an end-2-end understanding of our features while also producing artifacts needed to reproduce those drives. My key responsibilities included:
• Scope out and roadmap the tooling requirements and priorities
• Lead junior engineers on the team working on these efforts
• Support 5 VMS sister teams internally as well as various OEM partners externally by providing well designed and well tested visualization solutions.
• Design, develop and launch the various visualization solutions on the roadmap. The most recent one being developed is a tool for visualizing the hybrid powertrain optimization feature in vehicle.

Sep 2018 – Jul 2020 · 1 yr 10 mos

• Part of Google Maps team, with focus on the Vehicle Map Service (VMS). VMS is an in-car data exchange service that shares road and navigation data with the vehicle’s real-time operating system, allowing many vehicle components and systems to behave more intelligently as they gain awareness of the road around them. My contributions to the service include:
• Developed and successfully launched the feature complete version of the product for the first costumer ship to our OEM partner Volvo.
• As part of this, developed the pipeline for fetching, processing and publishing the detailed map data in the vicinity of the vehicle's current location.
• Designed and developed an in-memory cache to store the processed form of map tiles to optimize on resources and avoid repeated processing of the same area.

Aug 2016 – Sep 2018 · 2 yrs 1 mo

• Part of Firebase Notifications (FN) team. Firebase Notifications is an initiative to provide third-party app developers an easy system to send FCM notification messages to their users from a UI in the Firebase console. As part of the team, my major contributions include:
• Led the effort to re-design all the FN backend servers in order to provide the users a consistent experience across all other Firebase products.
• Designed, developed and led the major FN feature for allowing users to schedule recurring campaigns using FN APIs. In the past, the users could only send one-time campaigns.
• Designed and developed the FN backend support to provide A/B testing infrastructure to the firebase apps such that, an app developer could A/B test their campaigns using various targeting conditions with just a few clicks on the Firebase UI.

May 2015 – Jul 2015 · 2 mos · Menlo Park, Californie

• Worked on the Facebook’s Android App to provide users with more controls over their news feed in a much easier interface
• Developed the user friendly design for controls like reviewing the people, pages and groups one follows / unfollows.
• Conceptualized and developed the feature to “see-first” a few profiles so that you can always see their posts on top of news feed
• Structured and developed the “Discover New Pages” tab where users can discover page suggestions based on their profile
• Result : Impacting billions of users worldwide increasingly providing them with what interests them in their news feed and giving them more control over their feed thus increasing the average time spent by a user on facebook.

May 2014 – Jun 2014 · 1 mo · Greater Delhi Area

• Mentor: Prof Prem K Kalra
• Project Title: 3D Reconstruction of Human Body scan using 8 Kinects
• Worked on modules of sponsored project “Acquisition, Modelling and Display of Animated Objects”
• Generated human body scan using 8 kinects while establishing communication amongst two systems through TCP Protocol
• Calibrated the setup using Iterative-Closest-Point Algorithm further merging the 8 scans through affine transformation
• Results: A full-body stitched 3D scan of the subject was obtained