Brojogopal Sapui
Researcher working at the intersection of AI security, hardware security, edge intelligence, and trustworthy physical AI.
Short professional biography
I completed my PhD at Karlsruhe Institute of Technology with research focused on the physical security of emerging AI hardware accelerators. My work has covered side-channel analysis, fault injection, reliability, and trustworthy design for non-traditional AI hardware platforms including Hyperdimensional Computing, Spiking Neural Networks, and flexible neuromorphic circuits. I am currently working at NaMLab in Dresden, where my broader interest is cross-layer security that connects devices, circuits, architectures, accelerators, and deployed AI systems.
Core expertise areas
- AI/ML security across software and hardware
- Side-channel analysis and fault injection
- Edge AI and accelerator-oriented threat models
- Security of emerging AI hardware
- Trustworthy physical AI systems
- Cross-layer countermeasure design
- Hardware-aware robustness thinking
- Survey writing and research synthesis
Current technical direction
This website is organized around a broad but connected set of themes: software-side security of AI models, hardware-side security of deployed accelerators, cloud and edge deployment trust, and the security implications of newer paradigms such as generative AI, agentic AI, and physical AI. A central motivation is to explain how the threat model changes as AI systems move from isolated inference engines to interactive systems and then to embodied systems acting in the physical world.
A research portal, not only a profile page
The goal is to make this site useful both as a professional identity page and as a readable technical resource. Each topic page can function like a mini survey, while the research-watch pages can capture timely notes, open questions, and commentary on newly emerging directions.