Juliana Carroll

Title: Juliana Carroll: Inventor in Machine Learning-Driven System Architecture Automation.

Introduction

Juliana Carroll is an inventor based in Pinehurst, North Carolina, and is associated with Fannie Mae. Her work includes a patent for a system that uses machine learning and generative AI to automate aspects of system architecture.

Latest Patents

Carroll is named on 1 patent, titled “Machine learning-driven system architecture automation.” The invention describes a system that receives a user query, converts it into an embedding, and selects related artifact embeddings from a shared feature space. It then retrieves a knowledge graph whose nodes represent artifacts and whose edges capture semantic relationships. A generative model uses the query, selected embeddings, and graph as input to produce an output, which the system sends to the user’s device.

Career Highlights

Carroll’s patent reflects an approach to making architectural knowledge more accessible through semantic search and generative models. By combining query embeddings with a knowledge graph, the system is designed to connect user questions with relevant artifacts and their relationships.

Collaborations

Carroll’s listed coworkers include Subra Seshadri and Kiran Ramineni. The patent information provided identifies her association with Fannie Mae.

Conclusion

Juliana Carroll’s patented system brings together machine learning, knowledge graphs, and generative AI to support automated access to system architecture information.