May 2018 – Dec 2018 · 7 mos · Redmond Washington

Jun 2017 – Sep 2017 · 3 mos · Mountain View

Jan 2017 – Mar 2018 · 1 yr 2 mos · United States

• Experience:
• Winter 2018 - CSE101 (Algorithms), with Prof. Andrew Kahng
• Fall 2017 - CSE 202 (Algorithms), with Prof. Paturi
• Spring 2017 - CSE101 (undergraduate course for Algorithms), with Prof. Daniel Kane
• Winter 2017 - CSE 202 (graduate course for Algorithms), with Prof. Alexander Kulikov

Jul 2013 – Aug 2016 · 3 yrs 1 mo · Bengaluru Area, India

• Worked as a software engineer in a platform team responsible for developing L3 forwarding features for Arista EOS, a Linux-based network operating system for data-center and cloud network switches.
• Developed IPv6 Next-hop Groups support for MPLS and IP tunnels
• Designed and implemented IP-in-IP tunnel encapsulation support using IPv4 Next-hop Groups
• Implemented Unicast Reverse Path Forwarding (URPF) for 7050X series of switches in a team of 2 members
• Was involved in managing the major Q2’ 2015 Arista EOS product release; developed infrastructure tools for the ease of release management
• Developed infrastructure API in python over HLTAPI for automated testing by simulating switching protocols to test network switches efficiently

May 2012 – Jul 2012 · 2 mos · Bengaluru Area, India

• Worked as a Summer Intern in Data Protection Group.
• The aim of 2 month project was to to redirect the SnapDiff XML output to a file on NFS to replace HTTP to improve performance.
• We (a team of 2) designed and implemented an efficient method to transfer the diff of two snapshots of file system over the network by redirecting the output to a file on NFS which can be mounted and read by the client.

May 2011 – Jul 2011 · 2 mos · Kanpur Area, India

• Worked as a summer research intern under SURGE for 10 weeks.
• Research Problem: Task Scheduling in Heterogeneous Distributed Computer Systems
• Designed and implemented optimal scheduling of Matrix-Matrix Multiplication problem on heterogeneous system consisting of CPU & GPU to load balance sub-task across the components.
• Analyzed the variation in task execution time as a function of task division across CPU & GPU.
• Conclusion: The task execution time can be decreased substantially by distributing task optimally amongst various components of heterogeneous computer systems (CPU & GPU).