Vikram Varadarajan Kalkunte

Title: Innovations by Vikram Varadarajan Kalkunte

Introduction

Vikram Varadarajan Kalkunte is an accomplished inventor based in Pleasanton, CA. He has made significant contributions to the field of machine learning through his innovative patent. His work focuses on improving the efficiency of compute core workloads, which is crucial for advancing technology in this area.

Latest Patents

Vikram holds a patent titled "Just-in-time re-partitioning of feature maps for efficient balancing of compute core workloads." This patent outlines techniques for partitioning feature maps to enhance machine learning model processing. The method involves partitioning a feature map row-wise into multiple feature sub-maps, each defined by a split row based on the dense data element count. This approach ensures that each sub-map maintains the same column dimensionality as the original feature map. Furthermore, it assigns these sub-maps to various tensor compute units and memory units for parallel processing, thereby optimizing performance.

Career Highlights

Vikram is currently employed at Sandisk Technologies Inc., where he continues to innovate and contribute to advancements in technology. His expertise in machine learning and computational efficiency has positioned him as a valuable asset in his field.

Collaborations

Some of Vikram's notable coworkers include Kiran Kumar Gunnam and Matheus Almeida Ogleari. Their collaboration fosters a creative environment that encourages the development of cutting-edge technologies.

Conclusion

Vikram Varadarajan Kalkunte's contributions to machine learning through his patent demonstrate his commitment to innovation and efficiency. His work not only enhances computational processes but also paves the way for future advancements in technology.