About Me
I am a PhD student in the Electrical Engineering and Computer Science (EECS) department at Massachusetts Institute of Technology (MIT). I am fortunate to have Prof. Anantha Chandrakasan as my PhD advisor. My research interests lie broadly in creating energy-efficient circuits for AI and, in turn, utilizing GenAI for circuits, be it in generation, optimization, or performance prediction. I firmly believe that by creating better chips for AI, we enable more powerful and efficient models, which, in turn, support the development of the next generation of chips. This creates a continuous loop of improvement, where AI and chip technology advancements drive each other forward.
I graduated with a B.Tech and an M.Tech in Electrical Engineering from IIT Bombay, where I was awarded the Sharad Maloo Memorial Gold Medal for being the second-most outstanding student in terms of general proficiency, excellence in academic performance, extracurricular activities, and social services among all B.Tech/Dual Degree graduating students (1 in 999) at the 59th Convocation of IIT Bombay.
I was fortunate to have Prof. Souvik Mahapatra as my Master’s Thesis advisor. We worked on electrical characterization, modeling, and simulation of Stress-Induced Leakage Current (SILC), Positive Bias Temperature Instability (PBTI), and other reliability concerns in micro/nanoelectronic devices. Using a percolation model, we also modeled Time Dependent Dielectric Breakdown (TDDB) of very thin oxides.
I spent the summer of 2023 with the Memory IP team at Qualcomm in San Diego. We filed a patent on my work on making in-memory computing reliable.
I spent the summer of 2019 at Delft University of Technology (TU Delft), where I was hosted by Prof. Gerrit Bauer and Prof. Yaroslav Blanter at the Kavli Institute of Nanoscience. We worked on dipole-exchange spin waves for thin ferromagnetic films.
I have also worked with Prof. Animesh Kumar on an independent study project. We developed a method to obtain the time to failure distribution for any SRAM cell with any given RTN model.