About Me

I have always been fascinated by how electricity governs the world, from high-voltage power lines to nanometer-scale transistors. During my undergraduate studies at Virginia Tech, I explored these interests by pursuing everything from C++ programming to analog circuit design.

I found the perfect environment to explore these diverse interests during my Master's program with the Multifunctional Integrated Circuits and Systems (MICS) group. There, I led a pioneering project to develop a low-power wireless sensor node for livestock monitoring.

This was a foundational experience for me. Breaking new ground required involvement across the full stack, forcing me to think outside the box and master unfamiliar domains like PCB design and 3D printing. We successfully demonstrated that edge intelligence could be used to reduce power consumption by 50-fold, serving as a foundation for future work by our group in this area.

However, this work highlighted a critical bottleneck. I realized that deploying machine learning models on existing, general-purpose microcontrollers has a hard limit. To truly unleash the power of AI at the edge, we must co-design the algorithms and the hardware they run on.

This realization brought me to Purdue University to pursue my Ph.D. under Dr. Anand Raghunathan. As a member of the Integrated Systems Lab (ISL), I am currently researching Compute-in-Memory (CIM). Specifically, I investigate how to exploit Deep Neural Network (DNN) properties like sparsity to build more efficient hardware, and conversely, how to adapt these networks to better suit the underlying hardware.

Ultimately, this work allows me to pursue what I love most: tackling multifaceted problems that demand a broad perspective. Hardware-software co-design forces me to constantly expand my toolkit, and I look forward to continuing this work to enable the next generation of efficient AI.