This inventor holds 1 USPTO granted patent and 1 published patent application. Top assignee: Character Technologies Inc.. Active years: 2026.
Data source: USPTO (United States Patent and Trademark Office) public patent records. Weekly synchronization. How IDiyas builds this profile
Company Filing History:
Years Active: 2026
Title: Bowen Liang: Inventor in Large Language Model Inference
Introduction
Bowen Liang is an inventor based in Sunnyvale, California, whose patent work addresses efficient inference for large language models. His listed patent portfolio includes 1 patent.
Latest Patents
Liang’s patent, “Optimizing Key Value Cache for Large Language Model Inference,” describes processing an input sequence through transformer layers to generate tokens. The approach combines hybrid attention, multi-query attention, and cross-layer key-value sharing, then provides the generated tokens to the client device.
Career Highlights
Liang is associated with Character Technologies Inc. His patent focuses on optimizing key-value caching during large language model inference, a technical area relevant to the performance of transformer-based systems.
Collaborations
The provided information lists Noam Mordechai Shazeer and Myle Ott as Liang’s coworkers. Their shared association with Character Technologies Inc. places them within the same company context.
Conclusion
Bowen Liang’s patent highlights his work on improving large language model inference through attention techniques and key-value cache optimization.
Data source: USPTO (United States Patent and Trademark Office) public patent records. Weekly synchronization. How IDiyas builds this profile