About the Speaker

Mike Bartley is a semiconductor verification leader with over 35 years’ experience spanning software testing, formal methods, and Design Verification (DV). He began his career verifying safety-critical aerospace systems after completing a PhD in Mathematics, before moving into semiconductor DV in 1994. Mike has led world-class verification teams at STMicroelectronics, Infineon, Panasonic, and ClearSpeed, contributing to more than 15 processor verification programmes across automotive, mobile, aerospace, and AI domains. He founded and scaled a global DV services company to 450+ engineers, later acquired by Tessolve, where he served as SVP of VLSI. He now leads Alpinum Consulting, focusing on AI-driven semiconductor engineering and verification innovation.

Agentic AI in Design Verification

Overview

Design Verification (DV) remains the primary bottleneck in modern semiconductor development, consuming the majority of project effort and delaying time-to-market. This presentation explores how Agentic AI extends beyond traditional AI assistants by introducing autonomous, goal-driven agents capable of planning, reasoning, and executing multi-step verification tasks. Through practical examples including testbench generation, debug automation, coverage closure, and regression optimisation, attendees will see how AI agents can reduce manual effort, accelerate bug discovery, and improve verification efficiency. The session also examines multi-agent architectures, key adoption challenges, and a human-in-the-loop framework for safely integrating Agentic AI into production verification workflows.

Key Points

• Agentic AI transforms Design Verification (DV). It moves beyond traditional AI assistants by using autonomous, goal-driven agents that can plan, reason, and execute multi-step verification tasks, addressing DV as a major bottleneck in semiconductor development.
• Practical acceleration of verification workflows. Agentic AI can automate key DV activities such as testbench generation, debug processes, coverage closure, and regression optimisation—reducing manual effort while improving bug detection speed and overall efficiency.
• Safe adoption through structured frameworks. The approach explores multi-agent system architectures, adoption challenges, and emphasizes a human-in-the-loop model to ensure controlled, safe integration of Agentic AI into production verification environments