Scott Joynt

Title: Scott Joynt: Inventor Advancing AI and Machine-Vision Data Processing.

Introduction

Scott Joynt is an inventor based in Delaware, Ohio, and is associated with Insight Direct USA, Inc. His patent work addresses methods for coordinating specialized language models and preparing image data for machine-learning applications. He holds 2 patents.

Latest Patents

“Polling and Serial Prompting for Traversal of Multiple Specialist Language Models” describes a method for processing a compound prompt with multiple specialized large language models. The prompt is decomposed into a multistep plan, and selected models produce outputs for each step. An integration module then combines the step outputs into a coherent final response.

“Collection of Image Data for Use in Training a Machine-Learning Model” describes collecting and conditioning images to train a model to detect defects. Images of an item are captured from the same viewpoint under different wavelengths, including ultraviolet, infrared, and visible light. A defect identified in one image is marked at the corresponding location in the other images, helping create consistently labeled training data.

Career Highlights

Joynt’s listed inventions span two areas of applied technology: orchestrating outputs from specialist language models and organizing image data for machine-learning training. Both patents describe structured processes for combining information from multiple sources.

Collaborations

Joynt is associated with Insight Direct USA, Inc. The available information identifies the company but does not specify particular co-inventors or other collaborators for these patents.

Conclusion

Scott Joynt’s patent portfolio reflects work on practical methods for improving language-model workflows and preparing image data for machine-learning systems.