15 minutes with: Nikhil Antony

How would you describe your role to someone you’d never met before?

A lot of the work that I do is taking interesting concepts or technologies that are currently in research or at a really early-stage of development, and analysing how our clients could use them to solve their challenges and identify opportunities. This is often a mix of desk-based research and physical experimentation in the lab at our Global Innovation and Technology Centre. I work across markets, but predominantly in defence and security. That’s often where we see a lot of technology mature, to meet the need to develop areas of science for UK security.

One of my recent projects was looking at how various emerging technologies could be useful for a client, specifically in the area of sensing and processing. This involved looking at which external factors and organisations were driving the research, and working out where the client would need to act to develop the technology for potential use cases. The project was interesting because we assessed a wide range of promising research. The biggest benefit for the client was that we brought knowledge from across independent, often siloed areas of research and came up with novel concepts, such as using low latency computing with high precision sensing to create a product that could generate a completely new capability.

What makes PA stand out in our approach?

The big differentiator is our really wide range of expertise. There aren’t many companies where you find scientists, engineers, designers, strategists, and digital experts all working together.

That’s often why clients come to us. They’ll have big R&D teams that are very separated and don’t have the bandwidth to think about the problem at a higher, holistic level. What we do is get a group of people into a room and bounce different ideas around. I’m very focused on electronics, but I can easily talk to an industrial designer and figure out how the electronics and design can work together, ultimately creating a product that is considered from all disciplines.

Which projects have you worked on that have had the biggest impact?

I ended up learning a lot about the fabric industry even doing a deep dive into the effect of weaving patterns on fabrics. The specific type of engineered material has been a big topic in research for the last 10 years but has struggled to move beyond that. Partnering with the wearables team in our Dublin Studio, we found a way to make the specialised patterns on the fabric whilst also making it very easy to scrunch up and fold. We worked with university researchers to understand the physics behind the engineered material to ensure what we were doing was backed by scientific rigour.

This was for a higher frequency radio application, so we did very deep research, simulation, and modelling. This was needed as we were interested in topics that typical wearable electronics might not consider. For example, we looked at how the orientation of wool fibres will affect the speed of light. Then we moved to the lab, where we built experimental demonstrators to understand both the fabric and electromagnetic properties. We presented this work to key stakeholders who saw real value in the device in its ability to protect people and high value assets. Part of why it’s effective is that it’s a piece of fabric, so it’s very cheap, and you can fold it up and have a backpack full of them.

What are the biggest challenges that your clients are facing at the moment, and how are we helping to solve them?

One of the big challenges is often a lack of oversight of all of the innovation around a client’s product or area of research. They don’t always have complete knowledge about what’s happening in the wider world, and they’re trying to understand how different advancements, like ultra-low power computing, might impact the work they’re doing. When they come to us, we have that breathing room to help them engage in blue sky, big picture thinking.

What’s the biggest trend you’re seeing right now – it could be a methodology, or a technology, or a broader approach?

The big trend is making products that are agile, reprogrammable, and reconfigurable. This stems from the fact that it can take some time to get a product out into the market. By the time you’ve done all of the testing, the problem you are solving is likely to have changed, and the techniques may have progressed. So, there’s a real drive to try and make things that are reconfigurable. In the radio world, for example, we’re seeing more software defined radios that don’t have fixed electronics, and can instead be reprogrammed and used for different applications, such as Wi-Fi video streaming to long range telemetry.

What is the impact of advanced tech such as AI and quantum on your area of work?

Right now, the way AI is deployed is very cloud-based, in big data centres. We’ve spent a lot of time thinking about how we can use AI in a way that isn’t really slow or power-intensive. We’ve done a few projects around novel types of AI computing, and thinking about which problems they solve. If you’re streaming sensor data from, for example, a CCTV camera, a lot of sensitive data will have to be stored somewhere. If you can reduce the amount of raw data you’re processing and recording by putting a little bit of AI at the edge (next to the sensor), this delivers real-time insight to the user. But there’s a big challenge in making it low power and less resource-intensive while also making sure it genuinely adds something to the process.

There’s also the element of trustworthiness, which is something PA is looking at. We worked with a client on a novel type of AI which uses spiking neural networks (SNNs), which are low-power biological inspired neural networks.

Preventing these risks relies on solutions at multiple levels – the hardware we add to the chip, but also the training level. How do we train models so they are resilient in the field? How do we collect data in a way that doesn’t introduce bias or overfitting? This was quite an important step for the client because understanding the risks early on can speed up technology adoption and challenge the idea that new is risky.

Similarly, quantum could make a big difference to the way we do things. We have a quantum radio receiver in our lab that has a completely different trade space to normal metallic receivers. For example, while it can operate in more frequency bands than metallic antenna, it can’t be used to transmit signals. The way I see it, quantum isn’t going to replace what we already have, but it will sit beside it. It’s all about building architectures that enable you to put the right technology in the right place at the right time.

PA has looked at annealing machines which can be very good at optimising complex problems like dynamic timetabling. Optimising timetables is a problem that gets exponentially more intensive with classical computers. There are ways of translating the problem so it can be solved very, very quickly on a quantum computer. But, equally, for database querying where you have lots of small operations at massive parallel scale, you need a lot of data to flow through, which you won’t get at a quantum computing scale. Different problems will be better solved by different technologies, and complex problems may require a mix of technologies.

What advice would you give to someone who wanted to do what you do?

I started my career as an intern at PA, and when I’m talking to new PA interns, I often tell them to try to keep up with the research in their area of interest. At PA, you get a huge amount of independence in what you do and so being self-driven is really important. You need to know what’s out there. What I do is backed by a good understanding of the physics. That’s really important when you're dealing with complex, edge cases.

What are your professional and personal goals?

I’m pretty early on in my career, so my short- to medium-term goal is trying to find as many ways to learn about technologies that are coming out of research. Long-term, it’ll be really cool to work on something that I’ve come across in a research paper and help develop it into a product that provides a real-world benefit. I’m also understanding the benefit in building a bigger network. I’m quite technically focused, so I’d like to develop those softer skills of talking to people and trying to understand their problems, because that’s when you learn something new about a problem you’ve never heard of before and hopefully come up with a solution.