Company Filing History:
Years Active: 2025
Title: Souvik Kundu - Innovator in Machine Learning Efficiency
Introduction
Souvik Kundu is a prominent inventor based in Los Angeles, CA. He has made significant contributions to the field of machine learning, particularly in developing systems that enhance energy efficiency and data transfer. His innovative approach addresses the challenges faced by power-aware hardware in adapting to real-time changes.
Latest Patents
Souvik Kundu holds a patent for a "Machine learning model scaling system with energy efficient network data transfer for power aware hardware." This patent introduces a machine learning model swap (MLMS) framework that selects and interchanges machine learning models in an energy and communication-efficient manner. The framework is designed to adapt ML models to real-time changes in system constraints. It employs a similarity-based ML model selection process, which minimizes memory write operation overhead by selecting a replacement model with the most overlap in pre-trained parameters.
Career Highlights
Souvik Kundu is currently associated with Intel Corporation, where he continues to push the boundaries of machine learning technology. His work focuses on creating solutions that not only improve computational efficiency but also reduce energy consumption in hardware systems.
Collaborations
Souvik collaborates with talented individuals such as Daniel Cummings and Juan Pablo Munoz, contributing to a dynamic work environment that fosters innovation and creativity.
Conclusion
Souvik Kundu's contributions to machine learning and energy efficiency exemplify the impact of innovative thinking in technology. His work at Intel Corporation and his patented inventions are paving the way for more sustainable and efficient computing solutions.