2019 - now: Cloud Identity & Security - leading projects that enable seamless migration of enterprise workloads into GCP.2017 - 2019: Tech lead for 3 production critical services used by an overwhelming majority of google infrastructure (including GCP).

Developed a distributed data center load balancing solution for the Software Defined Networking (SDN) stack in Windows Server vNext. This project is a key part of the Windows Server MAS offering that was announced recently. Adapted parts of the distributed Software Load Balancer solution (Ananta) running in the Azure public cloud to integrate with windows server's SDN stack while developing this solution.My feature work and fixes span the full stack from REST API down to driver code.Key Projects include: - Live Migration of network virtualized tenant VMs sitting behind the software load balancer: TCP flows from clients both outside and inside the datacenter, survive live migration of workload VMs. - Ported some of Azure's load balancer control plane logic onto Windows Fabric, thus making it Highly Available. - Designed and implemented REST APIs to facilitate configuration of load balancer components by the data center administrator. - Designed and Implemented Diagnostics Infrastructure for the SDN stack. This included a REST API surface to correlate and bubble up common configuration issues to Data center and Tenant administrators. This dramatically lowered the support cost of our Solution.- Multiple Patents filed

I worked on the components providing conferencing services in the Lync server. Feature code and fixes I developed needed to work in a distributed environment and with high availability. Code shipped both on premise and to the O365 cloud. Key projects:- Lync Room Systems: Designed and implemented extensions to conferencing protocols (SIP/C3P) to accommodate room system devices in conferences, worked on provisioning of room system devices in Lync server and also contributed to implementing a web management portal for room system devices.- High Availability of Conferences: Implemented some parts of the conference data replication model, debugged and fixed numerous issues to make synchronous replication of conferencing data between front end servers, reliable.- Telemetry: Analyzed scenarios that needed better telemetry. Enabled upload of telemetry data to a big data store (Cosmos) for these scenarios and mined it regularly to debug service issues, measure service engagement etc.

Responsible for development of custom mini kernels for verification of hypervisor functionality. This position required the ability to debug systems software at multiple levels (hypervisor, kernel andoccasionally boot manager) and a thorough knowledge of x86/amd64 base systems architecture,virtualization extensions etc. I developed software components residing in a multitude of environments(Windows drivers, test mini kernels, hypervisor extensions and also Windows user mode components).

Responsible for design and development of software tools for component debug and design validationteams. This job required both an understanding of chip level hardware and embedded software design.

As a General Engineering (GE) Teaching Assistant at UT Austin, I designed and taught a supplementary class in introductory calculus to incoming engineering freshmen.

Developed a runtime verification tool under the application verifier framework. My work required meto have an in depth understanding of the Winsock API and internetworking protocols in general. I alsogained an insight into the working of some OS internals (Windows NT).

Designed and implemented an embedded system to automate collection of fault/maintenance datafrom the metro cabs running on the Delhi metro system. Supervised by Prof. D.V. Gadre (NSIT) & Mr.Jitendra Malik (DMRC).