Speaker: Nitin Kishore
Nitin brings 26 years of engineering and management experience, developing and bringing to market over 100 products. He founded Truechip in 2010, now providing one of the largest VIP portfolio. At Truechip, Nitin has worked with multiple customers across the globe. He founded TrueSilicon in 2024 with a mission to provide NOC Silicon IPs and other Silicon IPs to global chip designing companies. Prior to founding Truechip, Nitin was a Design Manager at Freescale Semiconductor (now NXP). He worked at FSL for 10 years. Prior to that Nitin worked at ControlNet, Goa. Hands on experience in frontend and backend SoC design flows.He holds a PGDM (Finance) from AIMA, and a BE (Electronics & Telecommunications) from Pune University, India.
Subject 1: Silicon That Learns: Can AI Replace Traditional Verification
Abstract:
This presentation explores how AI is revolutionizing chip verification while exposing its inherent boundaries. AI demonstrates extraordinary capabilities in
pattern matching, test automation, and defect identification—delivering tenfold efficiency gains for standard workflows—yet falls short in areas demanding human insight:
understanding design intent, imagining unprecedented failure scenarios, and making judgment calls. Through real-world examples, we examine AI's impressive achievements
alongside its costly blind spots. Mathematical proofs and practical concerns—including liability challenges, transparency mandates, and safety certification standards—
confirm that complete AI autonomy remains both theoretically unattainable and operationally risky. The solution emerges through human-AI partnership, where organizations
report 5-10x productivity multipliers while preserving the contextual understanding and responsibility inherent to human expertise. Verification's evolution points toward strategic
symbiosis: AI managing computational scale, humans supplying innovation, discernment, and ultimate ownership.
Target Audience: Semiconductor Professionals working in IP/SOC/ASICS projects