About

Vidushi Dadu is a highly skilled Software Engineer and Research Scientist with a passion for hardware-software codesign and designing programmable accelerators. Currently working at Google, Vidushi is focused on developing solutions to improve the overall utilization of an SoC by creating program idioms that support future applications and increase the lifecycle of a domain-specific architecture. Vidushi's research interests are centered around breaking the vicious cycle in the domain-specific accelerator design, where accelerators are designed for an application, then the applications are optimized for the architecture, and based on the feedback of not-directly supported applications, a new architecture is designed. Her work aims to support program idioms that are more likely to support future applications, systematically characterize prior accelerator efforts, and create general-purpose solutions that may be applied to various architectures including CPUs and GPUs. Prior to joining Google, Vidushi worked as a Research Intern at Microsoft, where she characterized Spark SQL on Azure Synapse to identify hotspot functions using flame graphs. She also developed a general methodology to profile at-scale workloads. As an Intern at Intel, Vidushi worked on developing a memory trace format to drive spatial architectures studies and demonstrated the utility of traces to find bottlenecks in the design. Vidushi holds a Doctor of Philosophy - PhD in Computer Architecture from the University of California, Los Angeles, and a Bachelor’s Degree in Electronics and Communications Engineering from the Indian Institute of Technology, Roorkee. Her technical skills include expertise in Software Engineering, Spark, SQL, and Azure. With 1.26 years of relevant experience, Vidushi is a highly skilled professional with a strong passion for her work, who is dedicated to creating innovative solutions that push the boundaries of what is possible in the field of hardware-software codesign and designing programmable accelerators.