What is it about?
Modern electronic systems, including smartphones, wireless communication networks, and data centers, rely on analog and radio-frequency circuits to interact with the physical world and process signals. Designing these circuits is a complex and time-consuming process that traditionally depends heavily on the experience and intuition of skilled engineers. Although artificial intelligence (AI) offers new opportunities to automate circuit design, relying on AI alone can overlook decades of established engineering knowledge and proven design techniques. This paper examines how AI can be combined with traditional circuit design methods to achieve more effective and reliable automation. It reviews existing approaches and identifies two major challenges: accurately predicting circuit performance without requiring extensive simulations, and ensuring that a circuit continues to perform well after its physical layout is created. To address these challenges, the paper discusses approaches that combine AI-generated mathematical models, established circuit design principles, and simulation-based verification. Building on these insights, the paper introduces CORES, a framework that brings together machine learning, AI-based reasoning, design exploration, coordination of automated tasks, and conventional optimization algorithms. Rather than replacing engineering expertise, the proposed approach aims to use AI alongside established methods to make analog and radio-frequency circuit design faster, more efficient, and more reliable.
Featured Image
Photo by Zach M on Unsplash
Why is it important?
Analog and radio-frequency circuits are essential to modern technologies, including wireless communication, computing, and electronic systems. However, their design remains highly dependent on specialized engineering expertise, making automation increasingly important as circuits become more complex and the demand for advanced electronics grows. Recent advances in artificial intelligence, particularly large language models, have created new opportunities for automating circuit design. However, AI-based methods often struggle to balance accuracy, computational efficiency, and reliability. This paper highlights why simply replacing traditional design techniques with AI is not sufficient and demonstrates the importance of combining AI with established engineering knowledge. The key contribution is the CORES framework, which provides a structured way to integrate different AI techniques with conventional design algorithms throughout the circuit design process. By identifying critical limitations in existing approaches and outlining ways to overcome them, this work offers a direction for developing more practical, scalable, and reliable design automation tools. Such advances could reduce design time, improve engineering productivity, and accelerate the development of increasingly complex electronic systems.
Perspectives
Through my research in analog and RF circuit design automation, I have seen both the potential and the limitations of artificial intelligence in solving complex engineering problems. While AI can accelerate many aspects of circuit design, I believe its greatest value comes from complementing established engineering principles rather than replacing them. One of the key motivations behind this paper is the need to bridge the gap between rapidly advancing AI capabilities and decades of accumulated circuit design knowledge. In my own research, I have explored how AI can help generate interpretable mathematical models, accelerate circuit optimization, and support automated physical design. These experiences have reinforced my view that successful automation requires a careful balance between computational efficiency, physical accuracy, and engineering insight. I see the CORES framework as a step toward bringing these complementary approaches together. My hope is that this perspective will encourage researchers to move beyond purely AI-driven solutions and develop more systematic, physically grounded, and practical tools for circuit design automation.
Subhadip Ghosh
University of Minnesota Twin Cities
Read the Original
This page is a summary of: Special Research Session: The Quest for the Optimal Use of AI in AMS/RF Design, July 2026, ACM (Association for Computing Machinery),
DOI: 10.1145/3770743.3807773.
You can read the full text:
Contributors
The following have contributed to this page
