San Jose, CA, United States of America

Bharadwaj Pudipeddi

This inventor holds 20 USPTO granted patents and 9 published patent applications and 1 EPO patent, primarily in Artificial Intelligence Optimization. Top assignees: Microsoft Technology Licensing, LLC, Intel Corporation, Bitmicro Networks, Inc.. Active years: 2008-2026.

IDiyas Innovation Intelligence. (2026). Inventor Profile: Bharadwaj Pudipeddi. Retrieved from https://idiyas.com/inventor/bharadwaj-pudipeddi

Data source: USPTO (United States Patent and Trademark Office) public patent records. Weekly synchronization. How IDiyas builds this profile

USPTO Granted Patents = 20 

% Patents Active = 75.0

 

Average Co-Inventor Count = 2.9

ph-index = 4

Forward Citations = 56(Granted Patents)


Location History:

  • Santa Clara, CA (US) (2008)
  • Campbell, CA (US) (2009)
  • San Jose, CA (US) (2010 - 2023)

Company Filing History:


Years Active: 2008-2026

Loading Chart...
Loading Chart...
Areas of Expertise:
Artificial Intelligence
Model Parallelism
Gradient Optimization
Memory Constraints
Deep Neural Networks
Error Recovery
Data Parallelism
Convolutional Neural Networks
FPGA
Distributed Systems
Latency Reduction
Thread Switching
20 patents (USPTO):Explore Patents

Title: **Bharadwaj Pudipeddi: Innovator in Hardware-Assisted Gradient Optimization**

Introduction

Bharadwaj Pudipeddi, residing in San Jose, CA, has made significant contributions to the field of technology with a remarkable portfolio of 18 patents. His innovative work primarily focuses on advancements in the optimization of neural network models, showcasing his expertise and dedication to enhancing computational efficiency.

Latest Patents

Among his latest inventions is the patent for "Hardware-assisted gradient optimization using streamed gradients." This patent describes systems and methods that facilitate optimized gradient calculation through a streamlined process. The innovative method involves a memory system designed to store weights associated with a neural network model consisting of multiple layers. During a single burst cycle, gradients from various workers are efficiently processed, including moving weights and writing back new weights generated by the gradient optimizer. Additionally, the new weights are transmitted back to each worker, thus ensuring a seamless flow of information and enhanced performance.

Career Highlights

Bharadwaj has held significant roles in renowned companies such as Microsoft Technology Licensing, LLC and Intel Corporation. His tenure at these leading technology firms has enriched his experience and expertise, allowing him to contribute to high-impact research and development initiatives.

Collaborations

Throughout his career, Bharadwaj has collaborated with notable professionals in the industry, including Marc Tremblay and Jinwen Xi. These collaborations not only highlight his ability to work in diverse teams but also emphasize his contributions in developing cutting-edge technologies.

Conclusion

Bharadwaj Pudipeddi exemplifies dedication and innovation in the realm of technology, particularly in hardware-assisted gradient optimization. His patents reflect his commitment to advancing neural network models and optimizing computational processes, making him a valuable asset to the technological community. With his extensive experience and collaboration with other experts, Bharadwaj continues to pave the way for future innovations.

Profile summary based on public USPTO records.
Data Sources: USPTO Patent Grant XML, Patent Center, EPO & CIPO • Normalized by IDiyas Innovation Graph. Methodology & provenance architecturePlease report any incorrect information to [email protected]
Loading…