Vedant G.

Working at Google Napa, Google’s exabyte scale data store for analytics. Backend engineer developing and designing components for supporting the data store for Google's Trusted Partner Cloud initative. Working on a 20% side project for making a graphical programming language.

Backend engineer for Google Stadia and AR, owned gamelet messaging and health monitoring systems. Designed flume pipelines for Kronos, YouTube’s video processing platform, for real time debugging.

• Worked on a crash analytics Python Kafka client running on Kubernetes • Designed Postgres backed REST APIs for crash related queries, including correlations based on signatures • Developed an iOS animated exercise app for testing accessory motion API • Created custom XML template language with features like inheritance for reducing Splunk code duplication

• Advanced in-house source-to-source compiler for optimizations and performing code path analysis • Designed and implemented algorithm for identifying memory leaks in the Java Card runtime at compile time • Developed automated compile-time code report generation and code path visualization • Implemented testing suite for testing function of tools and secure Java Card applets • Reviewed software source code to identify bugs and performance issues

• Built a low-cost indoor localization system to be used for interactive learning in classrooms of 20 to 30 students • Designed and implemented algorithm for tracking through reverse time difference of arrival • Facilitated concurrency, flexibility, versatility and replaceability through software design and in-house wireless protocols The link to the GitHub repository containing the code for ranging is here: https://github.com/VVHack/STM32F411E-DW1000

• Developed a spreadsheet driven motion simulation software within Siemens NX CAD software using NX Open .Net API • Worked on a VB .Net desktop application that controls servo motors based on user input

Developed a Labyrinth like video game and designed low cost wearable 3D printed gaming devices to play the game which was aimed towards the rehabilitation and recovery of stroke survivors. The device design involved a durable 3D printed movable device with cheap material and circuitry in order to appropriately translate the flex sensor output to a digital signal. The wearables were designed in a way such that they could adapt to any kind of body surface and still record joint rotation. I also designed complex 3D printed equipment and fixtures as useful attachments for existing medical devices.

Learned how to build and train state of the art neural networks and machine learning algorithms like Convolutional Neural Networks, Recurrent Neural Networks, General Adversarial Networks and Deep Reinforcement Learning algorithms. Obtained hands on experience through projects that involved dog breed classification, TV script generation, artificial face generation and teaching a quadcopter how to fly using deep neural networks. Projects were developed using Tensorflow and Keras in Python.