It’s been a very exciting start to the year for Cron AI as we welcomed several talented artificial intelligence and perception professionals to the team. Simon Tilbury has joined as our principal director, embedded systems & acceleration. Before joining us, Simon was the chief operating officer at Bones AI.
As an industry heavyweight in embedded systems, acceleration and productisation, Simon perfectly plugs the gap between our perception software deep learning and embedded hardware teams to help deliver a product to our end customers. His experience in development at Raytheon and Maxeler means he brings both a strong engineering background and the right processes to our engineering.
In this article, we enjoy getting to know Simon some more and hearing his thoughts on the industry and the opportunities ahead.
How did you get into this line of work?
I’ve always been interested in engineering. All through school, I really enjoyed and excelled in the STEM subjects. I was pretty good at English as well, but I was never arty — I never excelled with that as much. Both my parents and grandparents were also involved in engineering, so I guess I was destined to pursue a career in the industry.
After leaving school, I toyed with mechanical and electrical engineering disciplines, but ended up going down the electrical path. However, it became very clear to me that the thing I liked most was the computer side — the digital design and the software — rather than the analog electronics. So, I focused on specialisms for that, graduated and then quickly found work as a software engineer.
From there, I progressed through various layers of software engineering and then eventually up into team leadership and management. I then secured a role at a company which dealt with hardware acceleration involving FPGAs, which I guess is how the opportunity to work at CRON AI presented itself.
What do you like most about working in this industry?
The thing I like most about working in the industry is the fusion of creative and problem-solving skills. Outside of work, I enjoy wood and metal working, and making things which overcome a functional problem. To me, software and computing is an extension of that into the virtual world.
I am also energised by working with smart people. I love the mix of skills in the sector, where people are good at something but not necessarily the same thing. Every day, I work with very talented mathematicians, other engineers and software engineers on any given project, and the mix of disciplines and ideas is really exciting.
Have you had any role models or anyone you admired in the industry?
Not really within the computing or software industry but growing up my preferred book choice was always science fiction and fantasy, so I was really enthused by Douglas Adams, Terry Pratchett and the likes. And those authors tended to be early adopters of all sorts of computing. Douglas Adams was famous for being an Apple aficionado and used computers in the early days, even when they weren’t great to use. That sparked an interest in me, as did the things they were writing about. They were exploring concepts for futuristic computers, although in Pratchett’s case, they involved ants and various other crazy things. So I guess I don’t have any industry role models, but there’s definitely some from outside the sector who breached its edges.
Thinking about AI as a whole, what do you think are the biggest challenges that have been overcome?
In the wider world of AI, the biggest challenge I have seen is understanding how decisions are being made by AI. It is very easy to take a network and train it up by feeding it lots of data. But actually understanding how it is arriving at certain decisions is a real headache in terms of building a safety case and reliability arguments. I’ve definitely seen that as being one of the bigger challenges.
Another issue is the degree of data and computing involved. These are exploding at great rates to keep up with what is being asked of the technology. I think it is probably more the governance and understanding of what is happening that challenges the use of AI in that way.
What has been the biggest industry change you’ve experienced in your career?
I began my career in the industry a little over 20 years ago and when I started out software involved writing everything from scratch. You had libraries, but they were very simplistic and if you wanted something you had to build it. The idea of rapidly prototyping something actually meant writing the code, not just taking the code someone else had written and playing around with it. So that is a huge shift in terms of software and productivity and I think that is what’s driving the current boom.
In terms of hardware, when I started out processing and software ran on CPUs, and GPUs were used as graphics cards rather than these general-purpose machines. Additionally, FPGAs were used like DSPs for signal processing. The complexity of the applications which are now being powered by these devices was completely unthinkable at that point. But things have come so far that these are now possible, and it opens up whole new world in terms of what we can achieve.
Looking ahead, having recently joined the business, what do you think is the biggest opportunity for Cron AI now?
I think there is a massive amount of opportunities for Cron AI to capitalise on in the marketplace. Lidars aren’t a new technology, but 3D sensors are new in terms of how they are being applied, which means the ways people are using them is very novel. But that does mean that there’s not a great deal of support. Cameras are well understood, there’s so much out there you can plug into cameras to help you navigate the world of computer vision. However, it doesn’t currently exist for 3D sensors in the same way. So I believe we can harness these opportunities by making these sensors easy to use.