From OMSCS Coursework to Industry AI: Automating Wire Harness CAD Design

Abstract

Throughout my OMSCS program at the Georgia Institute of Technology, I have been developing a software system that automates the creation of wire harness components used in computer aided design (CAD) tools. Traditionally, as I have seen throughout my career in the automotive industry, engineers must manually model wire harness components in CAD tools, which is a time consuming and error prone process. My system applies artificial intelligence and data mining techniques learned across multiple OMSCS courses to transform this task into an intelligent automated workflow. This presentation describes how key OMSCS courses directly influenced the design and development of the system. Concepts from CS 6601: Artificial Intelligence, CS 7637: Knowledge-Based Artificial Intelligence, and CS 6750: Human Computer Interaction informed the architecture of the platform and helped shape a human centered design approach. Agile development practices supported rapid iteration and collaboration throughout the project. The talk also discusses the entrepreneurial and legal foundations that helped structure the business surrounding the technology. The resulting application integrates several modern AI models, including Anthropic Claude, Meta’s Segment Anything Model, and semantic search. The system functions as an intelligent agent capable of generating validated wire harness CAD components approximately thirty times faster than traditional manual methods while maintaining 98.81 percent accuracy. Attendees will gain insight into how OMSCS coursework can be applied to create meaningful real world innovations that combine artificial intelligence, human centered design, and entrepreneurship.

Bio

Image

Bryan Horvat is the founder and president of Amoeba Software, where he develops AI-driven systems that automate the creation and management of electrical wiring component libraries used in engineering CAD tools. He has over 13 years of experience in the automotive industry spanning wiring systems engineering, component library development, and engineering software. Bryan is completing a Master of Computer Science from the Georgia Institute of Technology through the OMSCS program in Spring 2026 and recently co-authored the SAE-published white paper “AI-Driven Data Consistency and Relationship Inference System for Agile Component Library Management.”