The increasing complexity of electronic system design requires significant engineering effort across schematic design, PCB layout, and FPGA development. Engineering teams often face recurring issues such as design errors, inefficient reuse of previous work, and time-consuming debugging cycles that lead to budget overruns and delayed delivery. At the same time, companies possess a large amount of valuable internal knowledge distributed across many experienced engineers. This project explores how AI agents can leverage such knowledge to improve design efficiency, ensure quality, and reduce development risks while maintaining strict data confidentiality.

The objective of this thesis is to design and evaluate a set of AI-based agents that support engineering teams throughout key stages of electronic system development: schematic design, PCB layout, and FPGA design. A master student will focus his/her thesis in one of the topics:

• For schematic design, the AI agent will assist engineers by identifying similarities with previously developed circuits, enabling reuse of validated design blocks, and recommending updated components where appropriate. Additionally, it will perform automated checks to detect inconsistencies and prevent common errors such as incorrect signal connections or copy-paste mistakes.
  
• In PCB layout, the focus will be on enforcing Design-for-Manufacturing (DfM) principles and ensuring compliance with internal quality standards. The agent will continuously review layout decisions, highlighting potential violations and suggesting improvements to enhance manufacturability and reliability.
  
• For FPGA design, the agent will target known development bottlenecks such as timing closure issues, clock-domain crossings, interface mismatches, and constraint definition errors. By providing guidance and automated diagnostics, it aims to accelerate development despite limited expert resources.
  

A key goal is to support engineers early in the design process by generating architectural proposals, identifying critical design considerations, and providing structured guidance.

The research will also investigate mechanisms for continuous learning from internal engineering expertise while ensuring strict protection of proprietary and customer-specific data. Techniques for secure knowledge integration, model updating, and privacy-preserving AI will be explored.

Ultimately, the project aims to demonstrate how AI agents can reduce development time, improve design quality, and minimize project overruns by addressing root causes such as schematic/layout errors and FPGA design challenges.