“Using AI to Speed Up Prototyping in the Lab: GaN Traction Inverter for an Axial-Flux Motor”

Generative AI assistants are becoming part of everyday engineering work, but how useful are they when a design must move from concept to a working prototype on the lab bench? This talk reports our experience using Claude, Anthropic’s AI assistant, to develop a GaN traction inverter for an axial-flux motor as part of a Ph.D. thesis within a Spanish national research project. Although the case study is an inverter and its multicore embedded control, the conclusions extend to other disciplines and very different problems.

Real laboratory examples show how Projects, Artifacts, Skills, Model Context Protocol (MCP) connectors, and the AI assistant integrated into the development environment can accelerate prototype bring-up. The talk also briefly covers specialized tools such as Claude Science and Claude Design, as well as Python-based user interfaces. Each example shows what the AI accelerated and how its output was checked on the bench.

It offers practical lessons for Ph.D. students and engineers who want to make these tools part of their professional work, without trusting them blindly.