Skip to content

Hello, I'm Leo

I'm Yonghao (Leo) Wang, an Associate Professor in Computer Science at Birmingham City University. This is my personal corner of the web: part research notebook, part writing desk.

If there is one thread running through my work, it is a simple question: how does information get from here to there — how fast, how efficiently, and can it be trusted when it arrives?

It started with latency

My PhD, at the Centre for Digital Music at Queen Mary University of London, was on low-latency audio processing. Latency sounds like a niche concern until you work in live broadcasting, where every millisecond between a microphone and a listener's ear is a design decision.

That background brought me naturally into broadcast engineering. I am currently leading the BEng (Hons) Broadcast and Communications Engineering degree apprenticeship, and I coordinate the broadcast technology area and related programmes in the BBC–BCU strategic partnership.

From live media to a different kind of network

The same obsession with latency led to Flexilink, a non-IP network architecture I have worked on for many years with John Grant. We originally designed it to carry live audio and video with guaranteed, deterministic timing, alongside ordinary best-effort data on the same link, without wasting bandwidth on protocol overhead. It is now standardised as open specifications through ETSI ISG NIN [1] [2].

Then something interesting happened. Because Flexilink does not rely on a routable IP fabric, and every flow is admitted and authenticated before any data moves, it turned out to answer a question that critical national infrastructure is now asking urgently: how do you connect energy, water and broadcast systems without exposing them to the whole internet? A design made for speed turned out to matter for security.

Which is how I ended up in cybersecurity

I came to security from networking and blockchain. I worked with colleagues on several university consultancy projects for blockchain and digital media companies. We used blockchain and steganography to protect media copyright and royalties. I was part of BCU's cybersecurity group for several years. Now my research focuses on the security and resilience of connected and cyber-physical systems, where networks, embedded devices and the physical world meet.

And now, AI and security at the edge

The newest strand is AI, seen from the hardware up rather than from the chatbot down. Working with an AI-chip start-up taught me how neural networks, including large language models, actually map onto silicon, and why energy efficiency decides whether intelligence can live in edge devices rather than distant data centres. AI now helps to defend systems, and it also helps to attack them. AI itself is not fully reliable either: it can hallucinate and make mistakes with confidence. Once AI runs inside infrastructure, securing it becomes the same old question about networks and trust, at a new scale.

Writing

Lately I have been writing more, often with AI as a thinking partner, about information, technology, and the places where engineering runs into physics and philosophy. The most recent piece, in Chinese, follows information from Han-dynasty beacon towers to the light cone:

Elsewhere

My friends and I organise the Technical Committee on Cyber Systems and Engineering in the IEEE Systems, Man, and Cybernetics Society. In 1948, Norbert Wiener wrote about control and communication in the animal and the machine, and Claude Shannon wrote his theory of communication. Since then, the relationship between people and machines has kept changing, and AI is changing it again. The committee looks at intelligent and secure cyber systems, from engineering practice to the questions behind it. You are welcome to follow our work and join our activities.

If you are a prospective PhD student or a potential collaborator interested in broadcast technology, non-IP or deterministic networking, cyber-physical security, or edge AI, I'd be glad to hear from you.

A little more about me, less formally: About.