My research interests lie in Operating Systems and Systems/Hardware Security. In collaboration with David Kohlbrenner, my current work looks at being able to exploit computational errors that are often silent. Cores that perform such silent, erroneous computations are called Mercurial Cores.
During Summer 2022, I interned at Microsoft Research , working with Monish Shah and Daniel Berger. As a part of my internship I evaluated various hotness tracking mechanisms for disaggregated memory infrastructures. My work impacted the ASIC design of future CXL memory controllers.
Built a tool that estimates fine grained energy
consumption of an application on per process basis.
I use hardware performance counter data such as
number of L1 cache misses, TLB misses, total number
of instructions etc. to predict energy consumption.
During my undergrad, I was a part of the UCSD SysNet Group, and was advised by Geoff Voelker. My project explored the benefits and overheads of using non volatile memory (NVM) for operating systems data structures. This project consituted my Undergraduate Thesis.
I worked with Sorin Lerner and William Griswold as a part of the Early Research Scholars
Program at UCSD. My project included crowdsourcing Software Verification
by gamifying the resource intensive problem of loop invariant
identification. My group studied the difference in user participation and
quality of data generated in the two versions of the game created for this
purpose and presented our results at the UCSD CSE Research Expo.
Checkout this video shot by UCTV featuring my project!