6 Questions with Peter Garraghan, CEO of Mindgard

Peter Garraghan, CEO and co-founder of Mindgard, is an internationally recognised expert in AI infrastructure and security. He has pioneered research innovations that were implemented globally by a leading technology company used by over 1 billion people. As a professor at Lancaster University, he has raised over €11.6 million in research funding and published over 60 scientific papers.

1. Can you give us an overview of your background and how you started Mindgard?

As a Professor of Computer Science at Lancaster University, I have a deep technical background in cybersecurity, AI, and Machine Learning systems. The journey that led to the

creation of Mindgard began back in 2016. I was leading my research team and started investigating the potential cybersecurity problems arising within AI and machine learning systems. It struck me that - as beneficial as AI could be - it also represented a massive potential attack surface that current cybersecurity tools and research were ill-equipped to handle.

I spent the next four years  conducting AI security R&D with a team of brilliant scientists and engineers. We had access to a big GPU data centre within my lab, which allowed us to really investigate these complex problems. Through this research, we confirmed that the challenges around securing AI systems were going to be immense as their adoption grew.

It became clear that, to properly address this, we needed to build an entirely new product around the technologies and insights we had developed through our research. Approximately two to three years ago, after making significant progress, I started laying the groundwork to spin this out into a company - raising initial capital, developing the technology into product demos, and getting the proposition in order.

Mindgard was officially formed in May 2022 as a spin-out from Lancaster University. For about 18-24 months prior to that, I was essentially working two full-time jobs - running my research lab and teaching at the university, while also handling all the startup activities needed to launch Mindgard. It was a lot of work, but driven by the belief that we needed to get this AI security technology out into the world ahead of the growing wave of threats against AI.

Since then, we've raised over £3 million in seed funding, built out our product platform, brought on a leadership team, and assembled a world-class group of AI security experts and researchers. My driving mission with Mindgard is to solve the very real cybersecurity issues surrounding AI that we've identified through years of research, and help enable the secure and robust deployment of AI technology across all sectors.

2. What is your favorite thing about the cyber security ecosystem right now?

What thrills me most is the opportunity to work alongside brilliant minds from diverse backgrounds to tackle this challenge head-on. At Mindgard, our team comprises top scientists and engineers from academic labs but grounded in practical application. We are leveraging years of groundbreaking research to develop specialised AI security solutions capable of making a real impact.

There is a contagious sense of urgency and purpose driving us. We understand that safely harnessing the power of AI could unlock transformative advances for businesses and societies globally. Conversely, the risks of widespread AI insecurity could be devastating on multiple fronts.

Fortunately, there is rapidly growing awareness and prioritisation of AI security. After years of my research team sounding the alarm about vulnerabilities in AI systems, organisations are now realising they can no longer treat AI cybersecurity as an afterthought.

3. What tips can you give to entrepreneurs considering starting a business within the cyber security space?

The most important thing for entrepreneurs embarking on a cybersecurity venture, especially one focused on securing AI systems, is to have a deep technical understanding of the field and clarity on the specific problems you are trying to solve. This space requires specialised expertise - there are no shortcuts to developing a solid grasp of the underlying mechanics and potential vulnerabilities of AI systems.  As any leading researcher in this space will tell you, AI security is by no means a solved problem, thus I would be wary of any startup claiming that they have all the answers, no matter what their marketing materials (and especially if they have no track record or even research publications within this space).

It's crucial to be driven by more than just money or hype. The best cybersecurity startups are founded by those with a genuine passion for protecting AI systems and a willingness to invest heavily in research and development to stay ahead of emerging threats. Shortcuts rarely lead to long-term success in such a rapidly evolving, technical domain.

The cyber risks facing AI systems are existential - both a massive business opportunity but also a critical societal need. Entrepreneurs addressing this space should be motivated by a desire to truly safeguard the secure development of AI as a driving force behind their commercial interests.

4. What cyber security tools or startups do you use on a regular basis?

We believe strongly in a defence-in-depth approach, utilising multiple layers of security controls to protect our systems and data. This includes robust access controls, encryption, network segmentation, threat monitoring, and regular security testing. We also stay on top of the latest threat intelligence to proactively identify and mitigate emerging risks. While the specific tools may vary, it's really the holistic strategy and vigilance that are key. Of course, all of our own proprietary platform development adheres to rigorous secure coding practices as well.

5. What's a great digital tool or hack you've discovered recently?

I've recently been using a password manager to streamline my online life. Instead of constantly resetting forgotten passwords or using the same one everywhere, a password manager allows you to securely store unique, complex passwords for every account. It's a small change but a huge timesaver and security boost.I’ve also learnt the appreciate the power of keeping sufficient slack within one’s calendar to focus on delivering and unexpected events – it is very easy to have an entirely stacked calendar, and then realise that a number of new activities require your immediate attention.

6. What can we expect to see coming out of Mindgard in the near future?

Mindgard has several exciting developments on the horizon. We recently launched Mindgard AI Security Labs, a free tool that allows engineers to penetration test their own AI, GenAI, and LLMss. This tool automates threat discovery, provides risk assessments quickly, and offers guidance on remediating model vulnerabilities models before deployment. Making this tool freely available is part of our mission to educate the AI community about the real threats to AI security.

We have a roadmap for additional product releases, including automated remediation, and additional integration with MLOps and cloud providers to be easily accessible a wide range of organisations.

In terms of thought leadership, I recently came back from speaking at NVIDIA GTC conference, one of the largest AI events globally. As one of the few cybersecurity experts invited to speak, I shared insights on separating fact from fiction in AI security. We also plan to publish two to three major research papers this year showcasing our latest advancements.

We are coming fresh off the announcement of raising £3 million in funding. To support our future growth we will be fundraising again, to accelerate expansion of our team and go-to-market efforts. We also recently onboarded a VP of Engineering from Snyk.

Overall, our focus is on providing organisations with the tools and expertise they need to securely adopt and deploy AI and harness its competitive advantages. With the most innovative products, top talent, and industry-leading research, Mindgard is well positioned to become the go-to platform for enterprise AI security. Expect to see a steady cadence of product releases, high-profile speaking engagements, impactful research, and continued team and customer growth in the coming year as we execute on our mission.