S2 E4: THE FUTURE OF TECHNOLOGY IS HUMAN AFTER ALL — with John Walsh, CTO, Fujitsu Ireland: Transcript

Where the Needle Lands
36 min readSep 24, 2021
John Walsh, Chief Technology Officer at Fujitsu Ireland

Listen to the podcast about John’s career in Fujitsu and the business of ICT from minutes 2 to 9.

(Geraldine) Fujitsu is 85 years old. You have been part of that family for 30 per cent of its span of time.

You’re one of the stalwarts, I imagine. So when you’re talking about customer relations, how much of that is to do with a relationship with you from your early days to your later days or is it just because of your role?

(John) That’s a question. Thank you for the — ‘you’ve been there 30 per cent of the time’.

(Geraldine) It’s just the way of looking at it.

(John) I didn’t feel old until now. I’ve been with Fujitsu for a long time.

A couple of things, in the context of Fujitsu and from my perspective, Fujitsu takes a very long-term view of things. By tradition, they’re a Japanese company. Their relationships with their customers tend to go back many years. That’s the first thing, right.

And there is a general theme in Fujitsu, which is about strategic thinking. In other words, look at it from a long-term perspective. They take a very long-term view of their customers and their relationships with their customers.

And the people at Fujitsu want to bring new things to their customers. So we have a pretty good partner network, an excellent laboratory. So we have the ability to bring stuff from the research phase, right up to the let’s-deliver-something-into-a-customer. So there is a very long term view taken at Fujitsu.

That’s not to say we don’t have to make a profit, and contracts and TCD and all those other wonderful things that we have to do, but there is that view taken is strategic. That’s the first thing. So that culture enables you to take time with a customer to do things.

Now, that shouldn’t be confused when we need to turn things around fast, and a customer wants to make a change. We have a lot of smart people. And they’ll have to change direction pretty quickly, but in general, what we do is we look at the strategic, long-term view of things with a customer, understand their business. Take time to understand their business because they themselves will have a vision.

They’ll have a five-year plan or a ten-year plan. In some cases, I’ve met customers that had a six-month plan. Get me through the next six months, John, and we’ll figure it out. Particularly in telco, if you work in the telco business, you’ll find yourself looking at a very kind of short-term view of the world.

So the company itself facilitates that strategic view of the universe. From a personal perspective, then. In terms of my position in Fujitsu. I’ve been lucky, let’s call it that every two years or so, I’ve been moved to a new project.

So in 25 years of doing work with Fujitsu, I’ve probably done the guts of maybe between 15 and 20 major projects. Some of those I was a participant in, in my early days. Others I’ve run top to bottom, soup to nuts. And from a technical leadership perspective. And the benefit of that is that if every two years you change your job, you get to see a huge amount.

You get to see a lot of technology. You get to exercise the company. You get to bend its muscles.

And when you’re working in an organisation with 128,000 people, there’s a lot of muscle. And let’s face it: there’s a lot of undiscovered muscle. And therefore, 15 and 20 major projects will facilitate you exercising that muscle.

And one thing by exercising the muscle is that you get to form relationships with people that you may not have been able to do previously. And like every other business, ICT it’s a people business, first and foremost, which facilitated me forming relationships throughout Fujitsu.

And those relationships also allow me to be massively effective. Because you get a situation in which if you hit a challenge with a customer or a problem with a customer, it is relatively easy for me to command the necessary resources I need to squash the problem effectively.

And then there’s the personality side of it. By my very nature, I get very focused very quickly on what I’m doing and almost to the exclusion of everything else to be truthful. And that is the thing that kind of facilitates me really driving down to the bottom of something very quickly.

(Mahima) I have two questions with respect to your focus. Do you meditate, or were you always this focused, or do you have a spiritual practice that helps you stay this focus?

(John) So that’s so funny you should ask me that. I’ve taken up meditation in the last two years as a way of de-stressing.

When you do this job, and the nature of the job is you’re continuously meeting customers. You’re staying on top of your technology brief. And right now, I have a responsibility to Europe, North and Western Europe, not just Ireland. And you are staying on top of your project delivery.

So there tends to be a huge amount going on all of the time. And one of the things that I find from a stress relief perspective, I learned how to meditate for 20 minutes. I’m sure there are people out there who could do this at the drop of a hat, but there’s nothing more difficult than focusing just on your breathing when the world is changing continuously around you for 20 minutes and thoughts coming in and going out continuously. So it’s taken me almost two years to learn that skill.

(Mahima) You had said ICT is a people business. What does ICD mean for you in the context of Fujitsu being in so many different industries?

(John) So that’s quite a difficult question to answer.

From my own experience, my experience has been in financial services, has been in telecom, has been in construction and manufacturing, has been in utilities. In those 15 to 20 projects that I’ve done, it’s been spread across multiple different verticals. So for me, you’re delivering, I called it ICT. You’re delivering service to a customer.

That service, in most cases, encapsulates technology, encapsulates service, encapsulates the management of the customer or the management of the customer’s customer. So it’s not just a matter of the good old ones and zeros, those things I find particularly easy.

It’s the, how do you wrap the thing in such a way that it’s going to work operationally? So from my perspective when I come to design something, I almost come at it from the backend. What’s it gonna look like in the real world, as opposed to theoretically, how are we going to build this thing? And, is it going to be a fine piece of technology?

My view is that very much tell me what this thing looks like 180 days in. I take a very operational view of how things actually meant to work. So when I talk about ICT, it’s not just purely the technical things I’m talking about here. There’s very much the technology, yes. And that’s important things have to be robust and have to be able to scale and have flexible enough to know you need to be able to change direction quickly. But it’s the wrapper. It’s the, how is it actually going to work in the real world? You need to be in a situation where you apply the technology, the service stack, whatever it is in such a way that it’s actually going to work for the customer.

Listen to the podcast about telecoms and the revolution of 5G and the Internet of things from minutes 9 to 19.

(Geraldine) Just from a customer’s perspective, you guys are, according to figures from 2018, and I know a lot’s changed since then, but you are in the top five largest IT service providers by annual revenue measurement.

If I’m trying to figure out as a company, my ICT needs for equipment and services, what’s the difference between you guys and IBM, Accenture, AWS?

(John) That is a massively different group of companies there. And effectively who are doing different things.

(Geraldine) But there are crossovers.

(John) There are, of course, crossovers. And I’m going to tell you what the differentiator is from my perspective.

So we have the ability to bring all of those companies together. The solutions that I design, I am not one of those people who believes in not-invented-here. If there was a piece of a relevant Fujitsu IP that makes sense that has been exercised previously and that will work for the customer, then, by all means, that is what I will choose.

But, if I need a platform of choice, and my platform will be AWS or Xor, from a Cloud perspective. And in some cases, Google. If I need an asset management system, I will go to IBM.

Have I ever done a piece of work with Accenture? I have, actually. I’ve done. Unfortunately, one of my claims to fame is the fixed point penalty system for Ireland.

And at the time, Accenture held the main contract in relation to, so I have worked alongside the Accenture, guys at Accenture.

From my perspective, what Fujitsu brings to the party is effectively the ability to be an able world-class systems integrator, and they are. You can ask for a solution, and you can be assured that what Fujitsu is going to do is they’re going to come to the party with the best from an integration perspective.

So again, it goes back to the point that I’ve made. If you’re designing something from scratch, start viewing the thing from an operational perspective, then figure out what the best components are. Who has the best components? Can you form a partnership with those companies and you bring those things to the table, and that would be classic of pretty much everything I’ve done in my career.

I’ll give you an example. In 2006, I was doing work with Telephonic in Dublin. They were O2, as they were at the time. And I had done small pieces of work for them on the legal side, believe it or not, in relation to capturing information around CDR’s — call data records. And the Garda wanted it, obviously, when they’re tracing somebody, looking for somebody, they go and produce a warrant, and you get their call data records.

So I’d start with a small project. But it was a complex project in the sense that there were features that were there from a security perspective, from a non-repudiation perspective. And complexities around the volume. Anybody who’s ever worked in telecoms will know that the volume of data which is produced in telecoms is phenomenal. There are whole industries built on analysis of telecoms data. This is the first job that I did. And the customer at the time was William Byrne. He was head of wholesale in O2 as it was at the time.

And William’s idea was that he wanted to build an MVNE. Everybody’s looking at him going MVNE? A mobile virtual network enabler.

So we had an opportunity to bid for this MVNE, and it was pretty much the first one in Europe. There was one in Japan at the time.

So mobile virtual network enabler is a virtual construct that sits on top of any operator. So how an operator works? An operator has a system centrally divided, typically between the business services system and the operational services system. The bottom half of that tends to be technical. The top half tends to be customer-focused. And what Willie was looking to do was effectively build a solution, which would be completely network and mastic. So you didn’t have to have all this tight integration between your billing system, your intelligent network, etc.

And he gave us the opportunity of coming in and discussing how we would build it. So I had three months to figure out, practically no experience in telco, three months to figure out how we would actually build a stack. And I got help from my colleagues in Japan. I got help from colleagues in Dublin. Some very good people on the technology side. And we put together a solution. That solution incorporated 16 different suppliers.

The reason I’m using this as an example is because it was Fujitsu’s approach to systems integration. We took 16 different suppliers, each of which had contributed something. So we needed an intelligent network. We needed a custom CRM system. We needed a customer billing system cause he wanted to do prepaid and postpaid. We needed a solution, which would do all of those things and we bolted it together.

And we went in and made the presentation to Willie and he said, I want to change a few things in this stock construct because I need to be able to do something slightly different. So he says, I want to stop the presentation. So this was about one o’clock on the day and everybody as a gas. They’re all sitting around in this room full of people. Maybe 50 people in there. So I want to stop. I just want to talk about it. So for the next three hours, all we did was talk about it. Now, it probably bored the living daylights out of everybody else in the room. Because everybody was ready for their presentation.

But that’s not what we did. We just talked about how we would put this thing together and how this thing would be a virtual construct and that would be able to sit at the top of anybody’s network.

So that was in 2006, 2007. To where I started in this conversation in terms of longevity, that customer has distributed with us for another five years. This is 2021.

And it highlights two things. One, it highlights the longevity thing that we talked about in the initial parts of this conversation. And the second thing is Fujitsu's ability to integrate and basically choose components necessary for the customer.

Do you want an example where that’s running live today?

So Giffgaff is one of the largest MBS in the UK. It’s sitting on this platform, it’s got three and a half million subscribers. Now you think about the number of subscribers in Ireland? So it’s equivalent to the size, effectively, of the Irish market from a subscriber perspective. And it’s run by 15 people sitting in the building on our site in Fujitsu in Dublin.

And it sits on top of Irish MNOs, mobile network operators and UK MNO’s. Most people are not aware that we don’t have our own network. What we have is a virtual construct that sits on top of everybody else’s network.

So we didn’t have to make the large capital investment from a network perspective.

(Geraldine) It’s like a form of piggy-backing off what’s already there.

(John) Yeah!

(Geraldine) This is really interesting. Teleco’s is a really good example of an industry that is either fading or it’s evolving. I’m not sure in its newer form and life, but what’s your view, John, on the diminishing of industries because of technology enablers like yourselves. Industries are fading or blurring and it’s about utilities and actually platforms?

(John) It’s a good question. The world is never going to stand still. It’s as simple as that. People are smart. People will come up with new ideas. Things are going to change and fundamentally you need to be able to adapt. There is no question. And even for Fujitsu, if you talked to me 20 years ago, 25 years ago when Fujitsu, we were at a very large hardware company, we did high-performance computing.

One of my first jobs was doing wind design with Airbus on the HPC. But things have moved on. We’ve all moved on. We have platforms that are effectively available in the cloud, they come with their full technology stacks. Some of them are more developed than others.

Customers are looking to move applications towards the cloud. The emphasis is much less on hardware. It’s much more around the software. And you can pretty much build anything with software nowadays. So things which were even hardware sacrosanct, I needed a network switch or I needed a storage device, it’s meaningless now in the sense that everything is focused on the software.

So the first thing I would say is you have to be able to adapt to your market. There are some fundamentals though, which will not change. And the fundamentals, which will not change. The need for networks and basic communication. And that really may be obvious, but it’s the evolution of networks you still need to have the basic infrastructure for carrying out networks.

And you talked about a telco and diminishing there. I think it’s probably the most exciting time for the telco. And that’s not because I’m having fun as far. It’s because I think people are coming to the realisation of the significance of networks.

So we are going to embark now on 5G. It hasn’t happened as fast as people wanted it to happen. But 5G for the user in the street, how much faster do you want your videos to be? How many more computer games do you want to play? That’s not of interest to me at all.

What is interesting to me at all is the extension of 5G to what they now call the E5G — the enterprise 5G. So enterprise 5G is going to be a sea change. We’ve talked about this, the Internet of things for five, six, seven years. But what is the true manifestation of the Internet of things? Can manifest itself until such time as you have the capacity to communicate with the edge, or the edge devices. The emphasis, therefore, has to be on the networks, the edge devices and the applications that go into those edge devices.

So you find yourself in a situation in which one level looks at a telco site, it’s a dive to the bottom, but in fact, what you need to do is look at it from a perspective of one part of your market may be in decline because of sheer competition and nothing else. But there’s another part of your market suddenly opens up. And there’s a hell of a lot more devices than there are people in the world. Therefore, the need to move towards E5G, IoT, and beyond that there’s the applications that run on top of those services. So those are basic services yeah?

Those are basic components that you need, but it’s really going to be done with the things like the applications.

Listen to the podcast about the difference between sophisticated networks and hyper-connectivity from minutes 19 to 26.

(Geraldine) You spoke about networks there, can you clarify the difference between sophisticated networks and hyper-connectivity?

(John) So in my world, right? From a hyper-connectivity perspective, everything is connected to everything else. I’m not quite sure if I see the point, to tell you the truth, right?

Why it’s absolutely necessary to do? Why does everything need to be connected to everything else? Even the sentence doesn’t make any sense to me.

I spent two years of my life at one point, working with Fujitsu laboratories on very smart networks. And the major focus there was on power generation, power transmission, smart metering. And there was this notion, even at that point, of hyper-connectivity of everything connected to everything else. But then you start looking at the real-world practical problems and the physics of the situation. You start looking at more importantly the security implications of what you’re talking about. And when you manage something centrally security, it is a relatively straightforward thing when you’re managing millions of things, which are distributed physically in the world, security becomes a much more complex thing to do. And I did spend two years of my life working with this stuff and I began to realise that we use phrases like hyper-connectivity without understanding the nature of the use case. So everything needs to come from what is the use case that we want to apply here.

E5G achieving connectivity from a use case perspective makes sense, achieving the security that’s necessary around that use case makes sense, but the concept of everything connected to everything else simply doesn’t make sense. Yes, you can expose data. And there’s a difference. You can expose data from those systems for consumption. Very different to the statement that I want everything connected to everything else.

I’ve been working on a solution for a particular customer. They are in manufacturing. They want to be able to monitor their production lines and they are looking at E5G as the solution to do it, which eminently makes sense. We will configure a solution for them. We will expose their data on what they call a head-end.

(Geraldine) Can you walk us through a more illustrated example of how the E5G is going to be used on their manufacturing line?

(John) I’ll typically walk through what it is.

If a manufacturing line goes down for a couple of minutes, it’s around 10,000 euros a minute. There are multiple manufacturing lines and in place today there are some monitoring of those manufacturing lines based upon serial communications. It’s old technology. It’s been around for 25, 30 years, it’s hardwired, it’s difficult to operate and it’s a very fixed limited amount of information that you get. And they tend to be aerobridged to the sense that it’s difficult to distribute information. So this particular customer that we’re working with right now we’re looking at a simple IoT system for them.

There are various different sensors on the manufacturing line. We are plugging those sensors into a new E5G network, effectively, a 5G-local network for that particular customer. We are providing what they call a head-end, which is a software front end to the data. So the data is inbound from the manufacturing line.

It arrives at the head-end, and then we slice and dice the data into what they call a time sequence database. Because obviously when you’re manufacturing stuff in real-time, everything has to be done as to a time sequence. That’s the fundamental construct. What then happens is once you have that data, you’re down in a position to start analysing what’s going on in that data.

So up until then, you’d say I’ve got a problem about a manufacturing line. Get me an engineer down there and fix something. Or there’s a compound mixing because when we do the end of a line analysis, there’s something not quite right. The molecular analysis that happens at the end, something not quite right. But when you get that data, it suddenly facilitates you having a much deeper understanding of your world and it allows you to distribute some of that data to your suppliers, for example. So that if there is a problem or there is a fault, it’s not only a matter of you knowing about it, but somebody who supplies something along your chain has the ability to know about it, pretty much close to real-time.

From my perspective, that’s fundamentally the difference between where we were ten, 15 years ago, serial analysis of manufacturing versus where we are today.

Are you hyper-connected? No, you’re not hyper-connected. You have a particular use case where you are taking the information, the E5G information from the various devices there, you’re exposing that information via a piece of application software, and you then have the ability to decide what you want to do with that data.

Whether you want to distribute that data to suppliers? Whether you want to analyse the data yourself? There’s a whole world now of supply chain, for example, where people are very keen to understand exactly what’s happened in the all way to the manufacturing process.

Particularly, where you have a complex manufacturing process where you have multiple suppliers into that manufacturing process. Where you have compliance involved in that manufacturing process and where your end customer wants to be able to track every single step that’s gone through from a quality assurance perspective. Then I suggest maybe that something like E5G and IoT is the way to go.

So consequently, circling all the way back to the initial question, which, in terms of the nature of business and some industries are in decline in that particular use case that we’ve just gone through from a networking perspective from a technical perspective, there’s a major opportunity. So it’s a matter of adapting to your market.

Listen to the podcast about creating human-centric computing systems and the safety and sustainability of society from minutes 26 to 32.

(Geraldine) And just on the power of ICT you’ve spoken about and how to harness such knowledge, and drive new value. For Fujitsu’s vision to create a human-centric intelligent society. John, how are you managing that and driving new value?

(John) What is the traditional view of computing? The traditional view of computing is you go and you meet your customer, the customer describes a business opportunity or a business challenge that they have.

And the very first thing you do is, typically, you write down your use case or your user story, or you start to learn about the process that the customer has to fulfil the challenge. And then typically what you always did was you always designed your system.

You’ll know in the entire conversation I just said there, we never touched a human person. We never touched on how a person would use the system. We never touched on how the system would be effectively operated, how people interact with systems, I didn’t mention that in those 15 seconds.

So that’s the traditional view of the universe. Now, that view of the universe clearly has been improved with agile techniques for doing development, where you’re much closer to your end-user. You have the ability to capture the stories. You do your review, you do your sprint, you come back to your end-user and you demonstrate what you’ve done.

But I would argue that it’s still very much process-driven. In the sense that there is a process that you’re going through from A to Z eventually irrespective of the methodology you used, you’ve arrived at Z. So fine, the difference between what I’ve just described and human-centric innovation is focused on the individual, which is unusual when you’re coming to developing large computer systems, but humans consume the output from large computer systems. So, therefore, doesn’t it make sense to start with the human and how they are interpreting the world.

And so consequently, what you will see from Fujitsu is what they call the human-centric approach. In that, what they want to understand is, ultimately, there will be a human consuming the system or this data or whatever it is, in what way are they going to consume it? In what way will it make the most sense that an individual has gone have to make some decisions on the basis of the information that they’re getting.

And again, it goes back to the point of how a system will actually operate as opposed to this theoretical and notional idea of how it will operate. Don’t build it and they will come, let them come and then build it. So that is I guess the most straightforward explanation that I can give of human-centric innovation and human-centric computing.

And it’s a big theme at Fujitsu in terms of how we design and develop systems and it manifests itself, how solutions are designed around people as opposed to people having to bend themselves into solutions to make it work.

(Geraldine) How does that play out in society then?

(John) So you will see if you go to the Fujitsu website the focus on the individual, you will see words that have real meaning in Fujitsu about the safety of society and sustainability of society.

And those are not fancy, corporate words that they use. They actually believe the stuff and the reason they believe it, if you think about it, for example, Japanese society, they’re in a part of the world where it’s subject to earthquakes, it’s subject to tsunamis. And something like the individual, the safety of the individual, the nature of where they live, the island population, the place has to be sustainable and energy-wise and all the rest of that. So the whole culture is around the safety, sustainability of society.

That’s where this is coming from at its core. And in terms of the way they design systems, if you go to Fujitsu laboratories, the things that you will see, like vehicle safety, you will see things like traffic management, a lot of it is very city-focused. A lot of it is very much about how an individual interacts with the city that it’s around them. So a lot of that focus is entirely on the individual and that has propagated into the culture of Fujitsu in terms of how they design the systems. And the focus on the individual, how the individual interacts with the world.

I’ll give you an example of human-centric innovation. I was the executive responsible for the commercialisation of a number of Fujitsu laboratory sponsored projects in Ireland, there were two major projects, which we have done under the auspices of Fujitsu laboratories in the last couple of years, the first one was in relation to the analysis of unstructured data using machine learning techniques.

Complex, multidimensional mathematics, tensors et cetera. And we use that in some elements of cancer research, where I may have an opportunity to talk about those a little bit later. The second one I’m perhaps most relevant was in relation to the interpretation of human sensor data. Now, I have spent a lifetime of developing complex projects for customers and complex computer systems, but this really brought home to me, the power of technology when combined with the human condition.

I had the opportunity to travel north and visit a gated community and that community for so where for some older folks, people living alone, people living in couples, et cetera. And those people allowed a number of sensors to be placed in their homes. We captured data, we curated the data. We analyze the data from those particular sensors.

I remember being there on my very first. And I noticed that there was a PIR - a passive infrared sensor in everybody’s hallway. Now, a passive infrared sensor, a simple sensor. You have them outside of your homes. They turn on and off the lights and winter. When I spoke to the SMEs on that project as to why a PIR was in everybody’s home, the PIR measured the number of times a person entered and left their home during the course of the day.

I was told there is a direct correlation between the positive or negative mental state of an individual and the frequency at which they leave their home. The greater the frequency, the more positive that individual is. They’re engaging with the world. They’re going outside into the fresh air. They’re engaging with other people.

Now we have multiple people working, on that particular project: mathematicians, engineers, data visualisation experts, psychologists, computer scientists. But what it effectively brought home to me is that when you combine all of those things with the human — look at what could effectively be achieved in terms of the actual individual.

Is there commercial gain? Yes, there is a commercial game. Irish population and the world population, particularly in the Western world, is ageing. The best outcome for those individuals is that they get to stay in the home for longer and they’re managed and looked after and kept healthy within their homes. And that for me is probably the best example that I have of human-centric innovation, where we take the person, the technologist, the SMEs, bring it all together and arrive at a human-centric solution.

It encapsulates both societies, humans and technology.

Listen to the podcast about hand recognition technology and machine learning from minutes 32 to 44.

(Geraldine) I have seen how you’ve been innovative to apply your crime surveillance technology to monitor and recognise complex hand movements. And this is an AI designs to check whether the subject completed the proper hand-washing procedure based on the guidelines issued by the WHO.

Now that’s brilliant, not for its initial reason, but it was the innovativeness around having a new application in the world.

(John) That’s an interesting thing to pick up on. And again, I’m surprised by your research. And I do know about it.

So interesting you should ask me something about this, given COVID and the times that we’re living in. I guess I made my fourth visit to the labs about 16 or 17 years ago, still as a relatively young engineer. And at that point, I was cognizant of Fujitsu’s work in video or engineering speak on structured data.

And of course, the time has passed and video analysis has been augmented with artificial intelligence and deep learning. And, this particular solution, in terms of the positioning of finger analysis, I probably saw in the labs ten, 12 years ago. What is the complexity around washing one’s hands?

Well, the first thing is digital finger analysis, where the position of fingers is not good enough. Why? Because one hand is actually covering the other when you’re washing your hands.

And secondly, you’re using some form of detergent, like soap. And the soap also blocks the camera’s view of the position of the hands. So what’s the answer?

Well, the answer is to combine the technologies. First thing is, Fujitsu has a deep learning engine, which is a shape engine, and that determines the position of one hand in relation to another. So as you wash your hands, your hands move one on top of the other, etc, and it learns those different positions.

The second engine is the motion engine. Why? Because we want to understand the number of repetitions when you’re cleaning your hands, the length to which one hand moves over the other. And if there’s any fuzziness in the image. So for example, if there was a shakiness in the hand, then you could exclude that from the video, etc.

Both of those engines are effectively brought together. And one engine feeds the other in order to derive a much more accurate result. So the information goes from the shape engine to the motion engine, and then back from the motion engine, back to the shape engine. And eventually, a very accurate result as to how well the hands have been washed comes out.

Now it’s an interesting piece of research work. I’m sure you can see that it has multiple applications and this is what Fujitsu tends to do. Again, it goes back to taking the long-term view. That is a piece of technology which can be applied to not only washing your hands but if you think about something in relation to food preparation. If you think about building a complex piece of engineering, like a very complex circuit board. If you think about the human-centric thing, performing surgery, or performing a very, very complex task in surgery.

You can take an apply this type of technology across all of those disciplines. It demonstrates a course Fujitsu’s depth in terms of the search and applying that search to real-world problems.

(Mahima) You just mentioned that you saw this technology 12 years ago. It’s coming into use right now. What does the technology that you’re seeing right now, or that we know will come into effect in another 10 years or what do you think the future looks like in terms of technology from your perspective?

(John) From my perspective, the future of technology - there is a huge amount of research work being done in terms of vision.

And I’ll go broader than that, on what I refer to as unstructured data. And that’s encapsulating sound, vision, text, language. Natural language processing will continue to develop, machine learning will continue to develop. I know people are concerned, let’s say, about what does that hold for us all?

But I worked on a data project in 16, 17, and we were using graph theory and tensors, which are fundamentally ways of describing geometry in multiple dimensions. And we were using that technology to read documents.

You might think, oh hey, the ability to read documents has existed forever and in computer terms. But the ability to read documents and extract meaning from those documents and context and semantics is still relatively new.

And we got to a point where, in that project, and I’m not an expert in oncology and cancers, but typically somebody who is a consultant in cancer has done 15 years of study on top of their medical degree to get to a point where they are the expert in what they do. Not only that, the really top people publish their research work. They have no end really of the amount of work they have to do to get to where they are. But what basically we did was we took 5,000 documents on cancer research and we assimilated that information on cancer research into a model. We are not experts and we had some people look at the research work in terms of different types of cancers, and proteins play a big part in the world when in cancer research. And we were able to do some of the mappings that would take somebody 15 years to learn in about three or four days.

And that was a combination of statistical analysis, the use of graph theory and specialist knowledge from people who are working with us to come up with these things. But for the very first time, what we could actually see was that we had the ability to train the machine to learn something which was really important and come up with some conclusions and most important in relation to the conclusions were that you could trace how it actually got to the conclusion. So the whole point, around adaptive reasoning, the whole point around inference, if you’ve spent six or seven years of your life doing the medical degree and 15 years of your life learning how to be a consultant, learning your trade and understanding and gathering that a really good person, a really good consultant is going to take all of that knowledge, all of the reading they have done and be able to work their way through the maze of knowledge to find the right drug to work on a particular cancer patient. Think about the ability for a machine to learn that material. To learn the same volume of material and to be able to arrive at a conclusion, which is similar.

And to be able to provide the evidence of how it managed to transverse that entire network of data to get to the same answer. People always say to me, when we’re working on these things, how opaque these things are. The real trick here is yes, you have to teach the machine how to do the things, but equally, the machine has to come back and say to you, this is how I derived that particular conclusion.

And that’s fundamental because that means that the real expert, the oncologist can then look at that and say that actually, this makes sense, or not. As the case may be.

(Geraldine) This is the powerful part of that equation that a lot of people are missing, that machine learning and all of that will do a huge amount of progress and work still comes back to somebody being able to guide it.

(John) Correct. I always say that we’re standing on the shoulders of giants. There’s probably never been a more exciting time in computing. Probably never in a more threatening time in computing either because of the way the nature of ICT is changing of course it is. But in my view, and I’ve been in this business probably 30 years now, it’s never been more exciting than it is right now.

And in terms of the advances that we are making. And dare I say it in terms of the advances that we’re making around humans. Around that whole human-centric thing that we started talking about earlier.

It’s really the fundamentals of we’re focused on the individual. We’re focused on the care of the individual and, I think you and I had this conversation before, but all ICT to me is just a set of tools. It’s just a combined set of tools to make the world a better place.

(Geraldine) That’s keeping a very grounded perspective on it because that’s all it is.

(John) And I’ve got the word technologist in my title and technology in my title, but fundamentally it’s a set of tools. That’s what it’s there for. It’s there to be used and ideally used in what they call the human-centric way at Fujitsu, but used for the benefit of people.

Listen to the podcast about mentoring and attracting tech talent from minutes 44 to 49.

(Mahima) What does it take to attract and retain tech talent in Ireland?

(John) Oh, that’s a really good question.

I have very strong views on this. I’ve less strong views on technology than I have on people. It’s learned over a long time.

So you know that people in Ireland have the ability to move, particularly if you’re qualified in mathematics, engineering, or a particular branch of physics, particular branches of engineering if you’re in computer science, computer applications, enterprise computing, there is no end to these things that you can do now.

If you’ve managed to graduate from one of those courses then you are going to be absolutely mobile in Ireland, in Europe, in the universe, it doesn’t really matter. So it is a challenge to attract new talent into the company. Occasionally, I’m asked to interview people who are senior architects and we always get a mix of folks from different backgrounds. So when I go interviewing those folks, the interview is almost in two halves.

There is get to know the person who’s across the table. I will take the technical knowledge as a given more often than not. I personally am looking for some traits. The traits I’m looking for is enthusiasm, which is the first trait I look for in an individual. The enthusiasm for the work and for the things that they’ve accomplished previously.

And I’m not just looking for this major computer science project. I did this and I did that. I’m looking for enthusiasm for life. For the things that are in their lives because enthusiastic people have the ability to overcome difficulty, serious problems, and obviously see the bright side of things.

That’s the first thing I always look for. And the second thing I look for is the fit in terms of the overall team. The projects that we work on tend to be so complex and so large and there is no question they get difficult. There is no question they get difficult. The person has to be able to work in the context of a team. They have to be able to work in the team and keep the entire team motivated, keep the entire team going in the right direction.

And I also look for individual traits. I don’t look for the same individual every single time. I’m looking for, I use this phrase diversity all the time. I look for diversity and the reason I look for it is because my mother used to say something like: “when the only tool that you have is a hammer, every problem looks like a nail”.

I go on the basis of, they have a different way of looking at a problem, you might tell them a scenario and say to them, what do you think of that? What would you do? And you will look for something different in the approach to the problem.

So that’s the nature of the first half of the interview. The second half of the interview is telling them about life at Fujitsu and telling them about the nature of the projects that we’re doing.

So I’ve lots of projects at my fingertips. Then you say, look, these are ten things I’m working on at the moment. Let me tell you about one or two of these things, which may be of interest to you. And normally in the course of the interview, if the person is genuinely interested in those things, you’ll find some common ground with them. And more often than not they’ll come to work for you. But that’s a tiny part of it.

The real part of it, for me, I learned this as a very young engineer, is mentoring. You have to make time for people. You have a very busy day. I’ve got whatever it is, two or three customer meetings today. I have a presentation to make to all the staff in the company at two o’clock which is on future technology directions of Fujitsu. And then I have two engagements this evening.

But you have to make time. So, at 10:30 this morning, I am meeting three of my engineers. And what that is — to discuss, is not about what are the current problems that we’re working on today, because every 15 minutes it’s a different problem. What, that’s the talk about is direction and careers and options open to them. It’s about growing the individuals and that’s something I learned as a very young engineer.

One of the jobs I had when I was back in the day, back in the early nineties was I went to work for ESBI, ESB International, which was a great three years of my life, really fabulous three years of my life. Because I was given lots of responsibility, I travelled to unusual places: St.Petersburg, just after the fall of the Soviet Union, Tashkent in Uzbekistan, Dubai, all these places. But one of the things that ESBI did was they mentored you. You were given a senior engineer to work alongside someone who was 30 years more experienced than I was. And I have kept that going all of the years that I’ve worked at Fujitsu. And particularly in the job that I’m now in, which is about mentoring the people who are working with you, giving them your insight. It doesn’t mean that you know everything, of course, you don’t know everything, but it gives them some grounding.

This is where I am in my career. This is why I am. These are the decisions that I’ve made, right or wrong to get to the place that I’m currently in. There is no doubt in your career you can do better than me.

I have a guy with me at the moment. He’s a brilliant guy, a head for detail, a great human being. My intention is that he will at a very minimum end up as a product leader in Fujitsu, which is he will own a whole section of the business. And he will develop and set the direction and trends. And I’ve spent literally every Friday morning at 8:30 with that guy, making sure that he’s on the right track.

So how do you retain people? You give them some insight into the company. You don’t oversell it. You have to be realistic about what you’re doing. You talk to them about their career, what their aspirations are. And without question, you mentor them. People I mentor tend to stay with the company, people that I didn’t, tend to go. So mentoring is a really important part of retaining staff as well as I am connected.

Listen to the podcast about digital transformation and John’s dinner party guests from minutes 49 to 54.

(Geraldine) It’s interesting then that the tagline of Fujitsu once used to be “Possibilities are infinite”, and now they say “Shaping tomorrow with you”. There’s a real change from we’re out there ahead to let’s co-create and learn together.

(John) That’s why people like me are in the position that I’m in, I guess. I say this all the time, I never claimed to be an expert in anything. There’s always something to be learned from the person who’s across the table.

You’ll learn something new in every interaction you have. The first statement there was based upon the infinity — “the possibilities are infinite”, that was based on the infinity symbol, which is still with Fujitsu, by the way. So that’s still there. But you’d see how the company has moved in its thinking.

We talk about this phrase, digital transformation and I’ve always defied anybody to describe what you mean with digital transformation. It’s a phrase that is branded around all of the time. But you cannot achieve a digital transformation with any customer, unless you have some appreciation of the customer’s business, understanding of the challenges that exist in the customer’s business, understanding all of the technology set that sits behind you and all of the suppliers that sit behind you and the partners that sit behind you. If you break all of those things to a single point, and in Fujitsu, they call it co-creation. But you’re bringing it to that co-creation point where all of a sudden you’re producing something new.

Because you have to use a technology phrase, you have multiple dip switches and multiple different configurations and this time around the dip switches are not set in exactly the same position as they were with the last conversation. And it’s about bringing it all to that focal point.

And it can’t do that unless you have their ecosystem, let’s call it, in place to make it happen. And if you want to know my definition of digital transformation, it’s about effectively creating something new. And you have to bring all of those things together and at one point, and then you’ll get something new.

(Geraldine) So for you, transformation is new rather than an evolutionary aspect? You know where I’m going with this?

(John) Yeah, yeah.

(Geraldine) The most elusive bloody definition.

(John) No, you can describe digital transformation as doing new things in new ways, but old things. Right?

(Geraldine) I’ll accept that.

(John) You can also describe that as evolution, couldn’t you? So you could also describe that fundamentally as I have ten combinations of solutions that I can use here. When I’ve used all ten of these combinations, I’m going to take these ten combinations plus add the 11th and 12th combination and provide the customer with something new. But purely on the basis of that, the customer has told me something new or told me something that I didn’t know. And therefore I develop something new out of it. So is it an evolution versus the creation of something new?

You and I had this conversation previously. Everything in the world of computing, to a larger extent, is a reinvention of things that were there before. There are some new things which come in, certainly, where we are in terms of development around AI, machine learning, natural language processing, yes. But in reality, it’s about combining the things that you have, the things you’ve learned, the solutions you’ve done before, the partners you’ve worked with, the suppliers you’ve worked with, combined it with one or two new things that you’re cognizant of from a technology perspective, or from a service perspective, or from a process perspective, and have a meaningful conversation with a customer where you could probably learn the new thing from the customer. And then combining all those things to create something. To evolve.

(Geraldine) I accept that.

(Mahima) So this is the question we ask all our guests. If you could have a one dinner party and you could invite three people alive, not dead.

(John) Do they all have to be alive?

(Geraldine) We know that you would have really preferred to go back there to Einstein and the likes, but no.

(Mahima) Three people alive.

(John) I was thinking about Newton.

(Geraldine) Newton even, sorry.

(John) There are fundamentals in the world of physics, which have been achieved by various different characters of which my greatest appreciation is for the likes of Newton.

(Geraldine) We’re sorry. We’re not going to luxuriate in that realm. You have to deal with the living.

(John) There’s your first problem.

So three people that I would choose from the living. Effectively it’s a dinner party. I would like Dalai Lama. The reason for that is it’s just that whole serene, calm, self-aware nature of that individual.

Let me think, the second person that I would choose and the person has to be absolutely living. I think I would like to meet Joe Biden. I think I’d like to meet Joe Biden.

I suspect he has a lot of stories to tell. He’s been around for a long time. His politics wouldn’t be a million miles away from my politics. I suspect I would like to spend a bit of time with Joe Biden.

Bizarrely, I don’t want to spend any time with any technologists, by the way, so you can take that as a given.

The third person I would bring to the party, I think is going to be John Nash. I’m not actually sure, John Nash is still with us, but John Nash was a brilliant mathematician from Princeton. And he worked on partial differential equations and gaming theory. And he’s the only man to have one both, the Nobel Prize for economic science and the Able Prize for mathematics.

Of course, everybody knows John Nash from the movie ‘Beautiful mind’ with Russell Crowe. But John Nash had some major, major hurdles to overcome in his life, including schizophrenia, which I think he was diagnosed in the 1950s with schizophrenia, but his situation improved from the mid-eighties onwards.

He’s an amazing mathematician., back when I was at the University of is very, very interested in the work that he did. And he overcame a lot of obstacles in his life. So I think he would make a wonderful third person at the dinner party.

What did I say it would be? Dalai Lama, Joe Biden and John Nash. How about that?

(Geraldine) The eclectic group

(John) That’d be a good dinner party.

(Mahima) Wouldn’t it just? Even Joe Biden has overcome so much.

(John) Correct. In his personal life. He’s overcome a huge amount in his personal life and I have nothing but admiration for the guy. I’m a big fan.

Do you know Bill Maher? You’re laughing. So Bill Maher does a weekly show in the US and he always finishes with a kind of a soliloquy and he was talking about Joe Biden. And he talked about the first hundred days of Joe Biden’s presidency was so dynamic and so action-oriented. In other words, real things were happening in the first hundred days. And the one acceptable bias people talk about nowadays is “they are older”. I will take strategy and I will take experience over youth any day of the week.