Ambrish Rawat

Ambrish Rawat is a senior research professional, specialising in AI Safety and Security and ensuring its responsible and trustworthy deployment. He leads efforts in red-teaming, security safeguards, and risk assesment, developing enterprise-grade solutions that help organisations confidently adopt AI by improving trust and reliability in deployed systems.

Ambrish has played a key role in enhancing the security of IBM’s flagship open-source AI models, Granite, and is a core contributor to the Trustworthy AI tool, Granite Guardian. His expertise spans Adversarial AI, AI Security, and AI Governance, with impactful contributions to open-source AI security, including Linux Foundation’s Adversarial Robustness Toolbox and IBM Federated Learning.

Since joining IBM in 2016, Ambrish has played a pivotal role in evolving AI security from a research-driven focus to tangible business outcomes. He has successfully integrated cutting-edge security capabilities into enterprise products, strengthening client trust in AI-powered solutions. As a cross-functional leader, he has driven strategic initiatives across IBM Research and product teams - bridging deep technical innovation with scalable real-world adoption. In addition to advancing open-source and product security, Ambrish has contributed to several EU-funded research programs and consulting engagements focused on secure and responsible AI. His work has been featured at premier AI and security venues including NeurIPS, AISTATS, ESORICS, and BlackHat USA.

Recognized as an IBM Master Inventor, Ambrish has made significant contributions to AI security patents. He holds an MPhil in Machine Learning and Machine Intelligence from the University of Cambridge and an MTech in Mathematics and Computing from IIT Delhi.

At the intersection of AI security, responsible AI, and business impact, Ambrish continues to drive trustworthy AI innovation, ensuring that enterprises can deploy AI solutions with confidence and reliability.