This inventor holds 1 USPTO granted patent and 2 published patent applications and 1 EPO patent. Top assignee: Mediatek Singapore Pte. Ltd.. 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: Kelvin Kae Wen Teh: Innovator in Large Language Model Technologies
Introduction
Kelvin Kae Wen Teh is an inventor based in Singapore, SG. He is known for his contributions to the field of large language models (LLMs). Although he currently holds no granted patents, his innovative ideas are reflected in his latest patent applications.
Latest Patent Applications
Kelvin's recent patent applications include the "Ring Buffer Storage Method" and "Ring Buffer Storage System Capable of Minimizing Extra Overhead Utilization." This method involves generating data from a large language model and writing it into a cache tensor buffer matrix. The process optimizes memory usage by shifting the starting memory address based on the number of input tokens. Additionally, he has developed an "Auto-regressive System and Auto-regressive Method for a Large Language Model." This method focuses on generating attention scores and outputs based on hidden states and encoded data, enhancing the performance of LLMs.
Conclusion
Kelvin Kae Wen Teh is an emerging inventor whose work in large language model technologies showcases his innovative spirit and potential impact on the field. His latest patent applications reflect a commitment to advancing technology in this area.
Data source: USPTO (United States Patent and Trademark Office) public patent records. Weekly synchronization. How IDiyas builds this profile