What was your background before you started Zenotech?

I did a PhD in applied mathematics in Adelaide, South Australia, and from there I spent two years as a postdoc in the NASA microgravity programme at the University of Delaware, solving dynamic equations for thin films – particularly the “contact line” problem, which concerns whether a liquid will wet a solid. After that, I joined the fluid dynamics group at BAE Systems, working at their technology centre in Bristol, UK. It was an interesting time to be involved in fluid dynamics because the aerospace industry was just starting to adopt codes for viscous computational fluid dynamics (CFD), going beyond the Euler equations for non-viscous flow. As I became more senior though, my job took me away from my natural inclinations. That’s part of why I left to co-found Zenotech: I wanted to return to the area I was passionate about, which was performing large Navier-Stokes fluid simulations on big computers.

What other factors spurred you to leave?

My co-founder Jamil Appa and I had talked about forming our own company roughly every three years since I started at BAE, but each time one of us would get promoted or something else exciting would come along. Then, a little over six years ago, BAE went through a round of redundancies, and we also both turned 40. That was a critical element in making us think “Well, it’s now or never.” It was a big step to leave that corporate environment. On day one, your inbox is empty. You have no contracts, no customers, no actual work to do. My wife told me to go outside and get a coffee, which was good advice because two months later, our inboxes were back to being full again, we were desperately trying to track down contracts from customers and the work was starting to pile up.

Who did you bring in to help you?

One of the first, and smartest, things we did was to join a local business incubator called SETSquared. It is very good at giving advice on how to go from doing technical work to running your own company, and becoming part of the incubator gave us a lot of contacts. We also buy in services such as accountants, lawyers, public relations specialists and designers when we need them. Design is especially important. In fact, it’s becoming the discriminating feature for a company offering any kind of online service. People’s expectations have been raised by the Internet and the various services they consume on it, and the user experience is critical. You cannot afford to have an out-of-date-looking website or an interface that’s difficult to use.

How did you get funding?

Zenotech has been “organically grown” – a sometimes pejorative term meaning that we are self-funded as opposed to taking venture capital money. We were offered money early on, but we were advised – I think quite wisely – not to take it at that stage. Instead, we have funded product development ourselves via consultancy work and collaborations with end-users in the aerospace, automotive and renewable-energy sectors. One of our products, for example, is a CFD code that is perfectly suited to doing aerodynamic design analysis on things like turbines and the airflow around buildings or over terrain. Wind-turbine manufacturers care a lot about how the wakes from one turbine impact on the performance of the ones behind it; similarly, the automotive sector wants to know where the vortices go if they’re designing a new car. We’ve developed that code over a number of projects, and a key feature is that it will run very efficiently on different types of hardware – central processing units (CPUs), graphics processing units (GPUs), all kinds of architectures.

You’re involved in cloud-based computing. How did that happen?

We offer a service called Elastic Private Interactive Cloud, or EPIC, and it started as a way of helping organizations use our CFD code without having to operate a supercomputer themselves or go directly to major cloud-computing vendors such as Amazon. The classic use case is a consultancy firm that has landed a big project that requires them to spin up hundreds, maybe thousands, of cores in a computing cluster over a short period – but then for the rest of the year they don’t need any computing power at all. It doesn’t make sense for a firm like that to go out and buy their own supercomputer, and they may also lack specialist IT people who know how to configure and run big computers in the cloud or solve the inevitable problems that arise. So we do that work for them, and then we help identify the cheapest available machine that will get their job done. You can do post-processing of the results in the cloud as well, and obtain the kind of “eye candy” images and videos that CFD simulations love to produce, without having to download tens or hundreds of gigabytes of raw data.

The development of cloud computing has also been interesting in a more general sense. Four or five years ago, when you talked to people about moving their computational engineering or high-performance computing applications into the cloud, security was the number one stumbling-block – it was used as a blanket excuse, really, for not doing it. There was this idea that because your data was being processed or your code was run off-site, it was somehow less secure, and that was a very difficult thing for cloud service providers to address. Now, though, the security argument has matured and people are far more cognizant about what they mean by a “security risk”. The relative merits of having an off-site, on-demand system where every access is logged, backed up, traceable and auditable, versus using a computer on your own premises with an open USB portal on the front and a regular traffic of visitors and students, have become clearer.

What have been your biggest challenges in developing Zenotech?

One challenge, especially if you don’t take investment money to start with, is that your salary suddenly depends on how much money the company has that month. Technically speaking, though, the challenges are the same as the opportunities. We started with a blank sheet – literally all our code has been developed from scratch within the company – and that means we probably made a few false starts before settling on a product line that works. Our initial idea was that we were just going to write a CFD code that runs on GPUs, but that rapidly turned into two product lines, and frankly it’s our cloud service that has been more widely applicable.

I also think we’re at a crucial point where both products are mature enough that I’d be willing to grow them significantly. That will take investment. The “organic growth” phase of the company is, I think, probably coming to an end. To work with larger businesses, it seems that a critical mass is needed, and that would involve growing the Zenotech team.

What do you know now that you wish you’d known when you started?

If I could go back and tell myself what to do, I’d have said to incorporate as soon as you can. Get a bank account running as soon as you can. In fact, get all that machinery in place because by the time you need, say, an accounting system, you’re probably already doing too much of that work manually. It’s far better to get processes in place and use a cloud-based system that costs you a little bit every month so that when the number of transactions does ramp up, you’ll be processing them efficiently.

The other thing I’d mention is that, coming from a large corporation, I always looked on the business world as being a wild and slightly scary place. There are a lot of unflattering images in the media of what business is like – shows like The Apprentice give you the idea that it is all about being nasty, ruthless and conniving – but in my experience people have been extremely supportive. There is a whole community who will bend over backwards to help and to give you advice if they can.

That segues nicely into my last question. What advice do you have for someone thinking of starting a company in this field?

There’s absolutely no substitute for giving it a go, but anyone looking to set up a technology business needs to assess their willingness to take risks. I had three kids under the age of 10 when I started Zenotech, and you need to be sure you’re not exposing yourself to a risk you’re not able to bear. Even with the best will in the world and the best product idea in the world, you have no idea what your competitors are doing or which way the wind is going to turn. Maybe your company is predicated on doing business with a firm that will cease to exist or a country that you won’t be able to trade with for some reason. Alternatively, sometimes the wind will blow in your favour.

A lot of people in the corporate world say things like, “You know, I’ve been at company XYZ for nearly 38 years now; maybe I should try starting one myself.” My response is, yes, you should – but only if you really want to. Don’t do it for any other reason, because you’ll find it’s quite hard at times. It turns up the volume on life when you’re running your own company. The highs are higher, but the lows (regardless of origin) are all yours.

• Enjoy the rest of the 2018 Physics World Focus on Computing in our digital magazine or via the Physics World app for any iOS or Android smartphone or tablet.