Ioannis Zacharioudakis

“Not all hardware faults are dangerous, but the few that are can change everything.”

Ioannis Zacharioudakis is a first‑year PhD researcher in computer science at Inria and the University of Rennes, France.

His work focuses on the reliability of emerging AI hardware accelerators in safety‑critical systems.

  • Industry relevance tags: AI hardware, Safety-critical systems, Autonomous systems, Space technology
  • Core research problem: How emerging AI hardware accelerators can be made reliable despite their sensitivity to hardware faults, and how to distinguish benign faults from those that can cause catastrophic failures in safety-critical applications.
“How reliable does AI hardware need to be when failure is not an option?”
Ioannis Zacharioudakis, The Short Version
  • University: Inria, University of Rennes
  • Location: Rennes, France

Ioannis Zacharioudakis is a first‑year PhD researcher studying the reliability of AI hardware accelerators.

His research shows that many hardware faults in neural networks are benign, while a small fraction can cause catastrophic failures.

He is deeply interested in DeepTech entrepreneurship and applying research in real‑world, safety‑critical industries.

Staying active, traveling, and exploring new cultures energize him outside of work.

Understanding Failure in AI Hardware

Ioannis’ doctoral research investigates the reliability of emerging AI hardware accelerators, systems that promise dramatic improvements in energy efficiency but come with new vulnerabilities. His work demonstrates that hardware faults do not affect neural networks uniformly.

Most faults are benign and leave predictions unchanged, but a small subset can lead to severe failures, especially in safety‑critical applications. This distinction is crucial for industries such as autonomous vehicles and space technologies, where hardware reliability directly impacts human safety.

From Research to Real‑World Impact

Beyond the technical challenges, Ioannis is fascinated by how academic research can move from the lab to the market. He actively follows DeepTech startups that scale energy‑efficient AI solutions and address sustainability challenges.

“I don’t want my PhD to stay only on paper. I want it to work in the real world.”

He is particularly interested in building bridges between academia and industry, and in learning how to pitch complex technical ideas to non‑technical audiences.

Building an Entrepreneurial Mindset

Through TESE Days, Ioannis hopes to develop the communication and leadership skills needed to transition from student to professional. Learning how to identify business opportunities within his research is a key motivation for joining the programme.

He values connections with industry professionals, mentors, and researchers from diverse cultural backgrounds who can challenge his perspective and help shape his next career steps.

Energy, Exploration, and Growth

Outside research, Ioannis draws energy from staying active outdoors and traveling. Experiencing different cultures reinforces his interest in working in international environments and tackling global technological challenges.

“Technology matters most when it survives contact with the real world.”