Aaron Phethean (00:15)
Welcome to today's show. We are joined by David Draper from Iris. David has had a fascinating journey from educator and teacher through to data science manager. Iris is a large software company with products in many different industries. And David shares his experience of building teams, building for innovation and quantum computing. So without a moment's further delay, let's dive in.

Aaron Phethean (00:40)
Today we're joined by David Draper from Iris. David, welcome to the show. I'm hugely excited to discuss quantum data and how things work over there. So welcome.

David Draper (00:46)
Hello.

Okay, you're welcome.

Aaron Phethean (00:55)
Let's dive straight in. One of the reasons we started the, you know, thought about having the conversation together was you've got this interest in quantum computing. Obviously you're working in a data role. Perhaps set the scene for us. How do those two things collide?

David Draper (01:11)
don't know, just probably my weird interest, I guess. I think my journey is quite interesting, I guess. I was a teacher for nine years, taught maths, computing, business, every single subject under the sun, like the puzzles we get hold of, and then kind of moved into sort of leadership in education and then providing analytics to schools. So it was kind of like, I don't know, it felt like a natural move. This is the next stage. I've always been interested in technology.

And I guess, you know, five to six years ago, I was looking at chatbots and how we could leverage data with chatbots in education, providing insights. And I think it's all around my, I don't know, I guess my teacher's, you know, inquisitiveness of trying to persuade people to use things. And I was like, okay, I could see quantum on the horizon. It's starting to pick up a bit of pace. It's interesting. bit,

Aaron Phethean (01:55)
Yeah.

I definitely agree that your journey does seem

fascinating. A teacher-type role through to technology seems like a pretty big leap. But on the other hand, I'm often surprised that they are fairly close to the cutting edge. They are doing new things. My kids are battling with AI and how it should be used as we speak. Every industry is doing the same.

David Draper (02:24)
Yeah,

I think it's one of the industries that, you know, got really impacted positively by COVID in a way. You know, I spent most of my life trying to get people to use Google Docs and Google Classroom, online Marking you know, I used it myself for a benefit and I did a Masters in Leadership in Education and it was all about using iPads and iPads in the classroom and it's purely because I hated marking I hated

those evenings sitting with a book and I was like, well, actually, if I could get the kids to mark themselves or peer assess themselves, or do multiple choice, just sort of look at their learning and think about how we can. Yeah. And I was like, I guess it's a bit of laziness. was like, actually, this is a bit of fun at first. And I was like, actually, this is quite interesting. I've always been into computers, into technology. How can I apply it to my job? And I guess it's the same for AI now as well.

Aaron Phethean (03:01)
takes out the whole marking bit.

You

David Draper (03:19)
How can it save me time? How can it save others time? We shouldn't be spending time writing loads of tests now, know, or even code boilerplate code is there for you already, you know, using, you know, using it every day.

Aaron Phethean (03:28)
Yeah, yeah, I agree. Yeah.

So then I suppose, we've sort of explored how you arrived in your role, but we haven't really explored who Iris are. They're quite a big company and do a few things. Why don't you just give us a couple of minutes on what they are, what they do and what you do for them.

David Draper (03:47)
Yeah, so Iris is, well, we're based mainly in the UK. We've got businesses that we've acquired in the US as well. We serve all around the world. We've got education products. education, MIS, finance, HR, payroll in the education sector, along with parental communication. And then we've got other domains. So we've got accounting domains. We've got an accounting product, which is our core product.

which is now a web-based and it's called Elements, serving accountants. And then we also have HR and payroll products.

Aaron Phethean (04:21)
Yeah, so a vast array of data and technologies and teams to deal with. What does it look like for the data team? What stage are they at?

David Draper (04:32)
So we've kind of got two separations in a data team. We've got an internal analytics team who are, you know, providing support, looking at sales and doing the internal. However, I sit on the product side. So what we're actually trying to do is deliver analytical solutions to education customers. I think that's also because of COVID, it's kind of helped that situation where people want to access data quickly. they want to see what attendance is like, in education. So we're actually on the product side, looking at analytics and how we can.

improve analytics, going back to customers and then utilizing it to make decisions.

Aaron Phethean (05:06)
What would you, so, you you see the team internally dealing with questions, probably quite ad hoc and, you know, I think many people watching sort of understand what that looks like. They're probably sitting in roles that work for a company. How would you characterize the, as a product side of a data team? What's different?

David Draper (05:27)
I think

it's, I think it's the requirements are slightly different. You know, it's still very similar, I think. I think it's working with product to work out what's actually needed. think real time analytics is quite key when it's serving a product. And I think that's also a challenge. Our products are quite historic in terms of, you know, the way they've been brought through the last 10, 15 years. it's, you're dealing with technologies that, you know,

not the latest technologies, you're not working with live streaming. So you've got to work out solutions to try and support that, but also support the product needs. It's quite an interesting problem to solve. It's a challenge, but it's exciting and working with leadership in education and leadership across the other domains. It's looking at how they can then utilize those, your data product to make decisions. So it's looking at how we can solve their problems, which I think it's quite exciting. It's, you you've got a

wide variety of people, wide variety of definitions of what they want to try and achieve. So it's coming up with a nice solution.

Aaron Phethean (06:27)
Yeah, yeah. I think that's a, that's quite an interesting point about the real time nature. I think when you're an internal team, this is my experience, you can explain that the reports happen the day after, or you can explain that, yeah, we haven't run that model yet. Or you've sort of got an opportunity to sort of say, well, this is, this is where we're up to. When it's the product, mean, expectations are much, much higher and they're like, well, it's in there. Why can't I immediately see it?

David Draper (06:44)
Yeah.

Yeah,

I think education again is quite unique as well. There's a vast amount of data actually. If you're about attendance, looking at a child's attendance, and that's a key area for everyone is children's attendance is falling down. It's in the news pretty much every day about some sort of attendance statistic. And that's really been impacted by COVID. But there was concerns during COVID about not seeing attendance. So we're...

We actually had to evolve. So if you're not working at IRS, working in my previous role, we had to evolve quite quickly. And we're actually looking at like Google classroom statistics, looking at how students were logging into the product. And actually we created what's, you know, it was called an at-risk register. So if a child was at risk, so basically we hadn't seen them online on virtual lessons. Was there a safeguarding risk? And that's what's quite interesting is actually some of these stats are quite key to.

Aaron Phethean (07:32)
interesting.

David Draper (07:49)
not just improving education, but finding students that are in need. So it's fascinating.

Aaron Phethean (07:55)
Yeah, I agree. And it

always helps when you've got a sort of a good grasp on the domain, you know, you can understand, you know, the meaning and the importance of a child not turning up and you know, the worth of that.

So I guess digging in a little bit further, we sort of discussed how the team looks, how the company goes about using its data. Maybe we could explore how you lead the team. How do you go about getting things done? mean, in a product role, I think I have an idea. I've been a product manager, product director for quite a long time, but each company is unique. So if you want to

start something or you want to build something, how does it work? How do you go about actually achieving?

David Draper (08:42)
So yeah,

so we follow kind of like an agile process. Um, and it's just to support us in terms of our journey. know what we're trying to achieve, particularly in the product. However, I think a lot of it is technology driven. What can we do there? You we use snowflake as a data warehouse and we found there's a lot of, you know, technology benefits to using snowflake, like data sharing. Um, you know, there's other nice elements to it. You know, we've got.

AI embedded. we're looking at how we can embed AI into our product. it's, you know, a lot of it is actually, you know, engineering going to product. So when actually we could potentially do this, there's a new feature out and I guess that's what's quite nice about this sector is it's evolving quite quickly. And, so there's a bit of, know, we can support product and making decisions, but generally we tend to work, you know, this is the roadmap, you know, it might change according to the technology, but you know, it's quite reactive as well, which is quite nice.

Aaron Phethean (09:40)
Yeah, I think that that does.

You know, characterize to me what product management is fundamentally about. a, it's a little bit of marketing. It's a little bit of educating. It's a little bit of Ford roadmap planning. It's a little bit of design and development, but you you can't build a product unless people understand what the scope or capabilities are. So, you know, you need to be at the sort of cutting edge of technology to say, well, yes, that idea is realistic or not. That's, that's, it's quite an interesting part of it.

David Draper (10:02)
Mm.

Yeah,

quite the thing is, think that's what quite interesting being on both sides of the fence. You know, I started as a product manager at Iris and now I'm on the engineering side. It's, know, you can POC anything. It doesn't matter what position you're in. You can create a little MVP POC, you know, I really like sort of like the Google design sprints, how they sort of iterate through crazy eights, that kind of, you know, we should all be doing that as product engineering people, we should be iterating through thinking about, you know, we could have

eight designs at start, eight ideas, but actually any one comes through. And it's, and that's quite, I quite like doing that. It's problem solving, isn't it? It's finding out what is a customer's problem. That's, let's solve it and let's solve it the best way. And you know, it might not be right the first time, but let's iterate and improve on it. That's what's fun. I guess fun.

Aaron Phethean (10:42)
Mm-hmm.

Yeah.

Yeah, yeah.

I think a lot of data leaders do find it fun. And I think one of the things I see is that often as they...

get into more leadership roles with more hands off. The fun can either drift away or it has to adapt and land somewhere else. How are you finding that sort of as you progress through your career and get more responsibility but less sort of code hands on day to day?

David Draper (11:25)
I don't know,

I guess my role is a mixture of both. So I do lead, but it's a lot quite still quite hands on. think in this industry, you still need to be quite hands on, no matter where you are. I met a few industry leaders at AWS a couple of weeks ago, and they were saying, they're still hands on and they're, know, they're much further on in their career. And that's it. That's why they like this, you know, particular area. They're having to continually learn, continually think about what is the next kind of improvement and then push that on.

through the levels and getting that through to product and engineering. it's nice. I think it's just a fascinating area. I think it's traditionally quite different, I guess. I think you do have to be hands on. There's news every day about a new model being released and new amazing features.

Aaron Phethean (12:09)
Yeah, I wonder if you keep

up unless you had the practical knowledge and the engineering understanding.

David Draper (12:15)
Yeah, hence, hence the I guess,

looking at quantum and thinking about actually what's the next bit. And that's what piqued my interest, you know, it felt like going back to basics as well, know, looking at quantum, you know, how do computers work, which is also fascinating.

Aaron Phethean (12:20)
Yeah.

that it's

definitely a good point. Let's explore that for a second. dying to dive into it. I think what would be interesting, I've been to a few quantum talks, I've been a little bit hands-on with it. Maybe back right up and

Try and explain to everyone what the opportunity of quantum is or what you foresee and the big shift being, if it were to become practical.

David Draper (12:57)
Well, I guess start from the basics, like how I think of it, trying to explain it to other people is, know, traditional classic computing, if you were looking through a maze, it looks through every opportunity to get to the end. And it does that one at a time. Whereas quantum computing is doing all of it at the same time, looking at all the routes at the same time. So it's not quicker in terms of getting A to B, it's going to A to B 20 times because there's 20 different routes. So that's the fascinating part for me.

and how I think, know, what's quite nice and hearing is that AI is going to be supporting the development of quantum computing. I think that's going to speed it up even more. You know, I Nvidia last week released, you know, they're creating a quantum stack not to compete with the quantum computing building, but actually to try and support, you know, the development and developing algorithms.

you know, created an infrastructure around quantum computing, which is quite interesting to see such a big business focused on AI actually supporting the next part of the technology. They must have, you know, being in a hardware, mainly an hardware company with, you know, they've got their own software, CUDA and everything, but it's how they're going to support that in the future when it comes to quantum computer. I think how they have seen it is they'll be able to support quantum computing rather than it taking their business away, which is.

Aaron Phethean (14:17)
Mm-hmm.

Yeah,

David Draper (14:18)
I found quite interesting, actually,

you know, that's where the development is. And I think it's going to speed it up. And I suppose, I kind of divert away the use cases and what I see. I forgot the guy's name... Mitchell, I read his book over a while ago, and it was all about quantum supremacy and what it's useful for. And, I guess being a teacher and all that, you know, bit fluffy, I'd like to

I hope it's used for good know, finding cues for cancer, you know, improving opportunities, but then you're going to have the, it going to improve my financial computing? Is it going to be able to predict those kinds of things?

Aaron Phethean (14:56)
do you make of the challenge around use for encryption, use for decryption, the possibility of...

it's use for accessing all the data that's been stored and encrypted and supposedly safe and then instantly is available to whoever has a copy of it. Is that a reality? Is that something that's already been thought about and battled against? Where is the world at on that front?

David Draper (15:10)
Yeah.

I think it's definitely a concern. I've read quite a lot of spy books, not quite, you know, very James Bond like books recently. a lot of it, a lot of the books that I've been reading just on the personal note, just the relaxing evening, there's, there's always some sort of quantum computing element where someone's broken the, you know, broken encryption, encryption, you know, and, know, it's caused disasters and secrets and stuff. And well, yeah, there is that element.

to it. And there's always going to be people trying to achieve that. And what's interesting, I recently, you know, watched a video and they were talking about the big banks and they're looking at, you know, encrypting their quantum telecommunications. So how and they're already doing that. So they're already thinking about how we go from bank A to bank B, no one can see what's being looked through by

twisting a cable and seeing the infrared or whatever, it's quite interesting to see that they're actually already thinking about that process. And you can't ignore the negative side. I suppose it's like AI, you can't ignore the negative side. People are using it negatively.

Aaron Phethean (16:31)
One of the interesting things that you said a moment ago, know, Nvidia, large company, you know, great valuation. Obviously, know, some tailwinds from AI, and yet they're already looking toward the next thing. I see that a lot of companies when it comes to AI investment.

kind of early in that journey as well, that they can see that AI is a disruptive technology. We might be able to see that quantum then becomes a disruptive technology. And all of a sudden, every company has to think about the risk to their business if it is as disruptive as everyone says. I wonder if you could, you you have any advice on how companies can deal with that tidal wave of change that just keeps coming and there will, I mean, AI right now, and there will be another.

David Draper (17:11)
you

Yeah

Aaron Phethean (17:20)
whether it's quantum or something else.

David Draper (17:23)
I think that

again, going back to COVID, I think it's been that rapid acceleration, people more accepting of technology, they had to use it. They were on Zoom calls, were, you know, I was, my wife was teaching online virtually, I was teaching training staff, 5000 staff across the trust, know, how to, you know, again, delivering those solutions. And it's, it's been prepared for it, you know, being ahead of the curve, having someone there who's, you know, advising you on what's next to come.

I think there's still a stumbling block. probably, you know, we talk regularly about data engineering. There's still a lot of stumbling blocks getting access to data. There is so many different technologies around there. You know, the data engineering landscape is changing quite rapidly as well. but it's being on pace, maybe not switching technologies all the time, but thinking about how you might implement it or, you know, being aware of it. you know, I've delivered training sessions, you know, just half an hour little.

snippets on what quantum computing might be able to do, how it works. Purely, I find it a bit fun and I wanted to share that I like going back into a little bit of maths, but I think in a couple of years time, I won't even need to, you've probably seen it build a quantum circuit. It's challenging. You don't really want to do that piece. You just want a library in Python just to, yeah, and I think that will come. It's not going to be long. There's a lot of libraries already out there.

Aaron Phethean (18:37)
value in

David Draper (18:45)
been involved in a couple of talks with Google and their search ML and search discussions about how they improve the technology. It's all open source. They're trying to develop it with people in that space. I purely joined as an interest having a look at it and thinking about, I imagine that being with ML and, you know, it may be in my space, it's not going to be used, but, you know, it'd be interesting to see what happens in, you know, drug discovery and that side.

Aaron Phethean (19:11)
Yeah.

David Draper (19:12)
you know, again, being a bit fluffy, all the nice and good pieces that would be nice to see what happens there.

Aaron Phethean (19:17)
Yeah, it strikes me that that particular role

in any company is quite valuable. know, the sort of find and educate. Maybe it's not for everyone, but you know, to have a few individuals who are continuously learning, continuously sharing, you know, actually the kind of source of, you you've mentioned a couple of examples we're doing with quantum, know, telling people how AI works, telling people about how data works. Actually sort of this sort of base knowledge can only improve with a little bit of, you know, internal education. And I see a lot of companies doing that really well and benefiting.

And the others that don't, I suppose, are at risk of just that wasted opportunity. There's probably someone there in the company who's willing and can then multiply the learning across everyone. Do you it in a structured way at Iris? Do you do it in an ad hoc way? Or how would you recommend for when you started?

David Draper (20:08)
Yeah. we,

so we actually have a kind of like architecture community where we, it's actually delivered by one of my colleagues and people sign up to do half an hour talks. So you do it in something that you've learned or you're interested in. So again, mine is all in machine learning, quantum data. I'm actually delivering one next week on how everyone should be learning Python. you know, very traditionally C sharp.

dotnet house. So it's, know, you know, why should everyone learn Python? And, you know, think about the future. It's a bit of bit

Aaron Phethean (20:41)
It sounds like a great topic. What would

you make of vibe coding if you're on the topic of learning new languages?

David Draper (20:47)
I, yeah,

I know. I've thought I see that it's, don't know. I don't know. I'm always very much. like having a look at these things and I evaluate it. And it's the reason I love Python because I wasn't a coder, I guess, when I first started, I, you know, taught a little bit of it and played around with it. And I found it easy. I always thought, you know, data science, data engineering, that's the place I want to go. That's what I want to learn. But again, that's what's quite nice about AI now.

Aaron Phethean (21:12)
Yeah, yeah.

David Draper (21:16)
actually can teach you other areas of coding. you know, probably again, another day of engineering, you know, people are looking at Scala and let's pick up Scala. It's very much Python like, let's, let's get, AI to help teach me how to implement it.

Aaron Phethean (21:26)
Yeah.

think that is the... yeah, obviously there's Cursor taken off.

you know, it's an assistant to coding, you know, the kind of idea of Vibe coding, you're not necessarily having to code, but you're coming up with the ideas. You know, there is a huge value and there will continue to be a huge value in understanding and then being able to apply it. you know, knowing that there are limitations of a network, knowing that there are limitations of storage or computing and then obviously being able to deal with those and turn that into something useful.

whether it's a new language Python or AI assistant, you're helping code in something new or just application builders that are purely AI driven. I think the understanding element will continue to be hugely valuable.

David Draper (22:15)
Yeah, how I've kind of explained it to a couple of people, like if you look at, you know, not the real definition of accuracy, but you know, lot of it's, I've tried to do it in old money, you know, if you look at it, education, ABCD grade, I said it's like a B grade at the moment. It's acceptable, but it's not quite an A grade or a star grade. It's not, it's not perfect. You know, if I'm going to write a piece of code and you know, it's going to take me a long time. What it helps me do is

build that boilerplate code out and it might not be perfect, but I can then go in because of my knowledge. I can go in and change it and adapt it to how I need or extend it out. It just helps me with the basic tasks that should be that boilerplate code. can just deliver that. Yeah. Yeah. You know, if it's, I use it as a bit of an assistant as well, you know, so I don't have to go and ask someone else. If there's something that I'm finding a bit difficult, I've not used a particular technology.

Aaron Phethean (22:57)
The time consuming parts for sure.

David Draper (23:08)
I go and ask it to explain it or I put the documentation in and say, can you explain this documentation or write me a piece of code to show me how this documentation works? I've done that and it, and it helps me. It's like a nice little advisor, but again, it's not perfect. Um, but it has helped me certainly. Yeah.

Aaron Phethean (23:21)
Yeah. And then you can tweak the rest of it. I mean, I

definitely see it's used a lot in the kind of engineering, let's say infrastructure, know, the kind of, you know, almost the...

to the mechanical part of the software and data stack. What about the analytical part? How do you see the maturity of it for answering questions or producing queries that are valuable? Are you using it on that front?

David Draper (23:52)
Yeah, so we're looking into it at the moment. I feel like there's a value just, again, education background, trying to teach people how to use data and look at an interpret graph is a challenge. You know, it's not everyone's cup of tea, people are interested in it sometimes, but visually, it's visually appealing. And it's sometimes actually you miss something in a graph, you miss that tiny little bit of detail. And we've, you know, we've checked a few graphs in and it's, you know, giving you a bullet point list of what it might show.

But again, it's that advisory piece. It's not 100 % perfect. Don't take it as a written rule. You need to interpret it yourself that this might be a good guidance. So that's where I see it benefiting.

Aaron Phethean (24:27)
I do hear

that a lot around called storytelling and the sort of art of being in data and telling the story. mean, that is the kind of hugely valuable bit of an analytics team is yes, everyone can see the data, but not everyone can make sense of it. And to have someone that actually helps interpret it and then explain what it's showing and then debate about it. It might not, the data and the chosen chart and the way you've sliced it might not be

David Draper (24:35)
Yeah.

Aaron Phethean (24:55)
really what's happening, but you can have the debate about it with the information in front of you. That definitely feels like, again, there's a bit of an engine AI assistant part, but it's not going to tell the story necessarily. just generates the idea.

David Draper (25:08)
No, it's,

but like it's, again, go back to another education example. You know, there's a bit of a in joke when the weather's bad, behavior is really bad in a school. And I was like, well, actually, I'm in a, I'm in a position to improve it. Oh, sorry, prove it. So I was like, right, fine. I'm going to get some weather data and match it against behavior data. And there was a clear correlation. But then when you think about it and go into a bit more detail about it, it's actually, well, kids are going to go inside.

They don't want to be outside in the rain. They don't have umbrellas. Half of them don't want to wear coats because it's not cool. So, and they need to run around. You know, I was one of those kids on, you know, always wanted to be outside. I hated being inside and would end up running around. just, you know, it's that then you've got that explanation. you know, it's data on a page. It's got a small explanation, but there's actually a story behind it. And thinking about actually, there must be a reason. Can we find a reason behind this?

Aaron Phethean (25:37)
Yeah, not going to burn off steam.

David Draper (26:03)
what this graph is showing has our revenue dropped, but is there a reason? it, is it something to worry about? Is it not to worry about? And that's, guess where, you know,

Aaron Phethean (26:06)
Yeah.

And what is the

cause to be the classic? You you correlated it, but you could have a different situation where they were stuck indoors, forced to do something that they don't normally do or want to do. It wasn't caused by the weather, but it was caused by the situation.

David Draper (26:25)
Yeah.

Yeah,

I actually explored this for my dissertation in science was actually so, you you're looking at predictive model, and it's giving you a prediction. And I guess it gives you like a percentage prediction. But what's the reason behind that prediction? What you know, it's model explainability, it could be one little thing, you know, I don't know, they forgot to turn up on time, just a random example. But it could be one little thing.

Aaron Phethean (26:34)
Give it.

David Draper (26:57)
and it could impact the whole model. So is it significant enough? Like that prediction might say, yeah, that's definitely going to happen. But it could be one little minus significant thing that actually, you know, from being an expert isn't relative to that prediction, and you can ignore it and or have it, you know, you might have a conversation. It was probably quite deep, if you're looking at one line of a long list of data, but when you talk about students, or talking about staff in HR or payroll, it might be significant because

Aaron Phethean (27:25)
Yeah, yeah, yeah.

David Draper (27:25)
know, member

of staff might leave or I guess in our products, yeah.

Aaron Phethean (27:28)
and then it's been impacted immediately.

So David, we've discussed quite a few different things, you know, technology and, and your career path. And, you know, one of the kind of things on my mind is, you know, everyone's journey is a little bit different and we learn from different people and we have different influences. I wonder if you might share some of your key people in your life, like who influenced you and helped you and what maybe as, as advice, like what should people look out for?

to take their next career step or even look out for when they're employing someone.

David Draper (28:04)
I it's in terms of people that have influenced me, I've had a few, but I felt like I've always been in good teams, but it's not just particularly one person. You know, looking at my role with a large group of trust, which was my role before Iris, we had a really good team around us and they were all had different experiences and they were all older than me. They're all more experienced in education data. Being a teacher, you're looking at specific class data. So it's like a different perspective.

I always felt leaning on a team is really good and building a nice team around you. And I have that iris as well. There's a lot of experts. There's a lot of people that know more than me, but I, you know, ask some questions and they asked me questions and it's like working as a big team and not, they might not specifically be in my team, but you know, they're around this, around the outside and yeah. And sickly engineering, I find there's a lot of people that are really keen to support each other. And that's what's nice. I think it's just.

Aaron Phethean (28:50)
of extended.

David Draper (28:59)
building a nice team around you and having a lot of people to lean on. You know, I guess I've had people individually as well, but it's, you know, I think in this new space as well, it's a challenge, but you know, I've got a, we've got a VP of data science who's not in my team, but every now and again, I meet him and we just have a general conversation about what, what he's doing, what I'm thinking about doing, you know, you know, what's just been released, just something exciting, just to pick up on.

Aaron Phethean (29:18)
Yeah.

Yeah, just on the level, yeah.

David Draper (29:25)
Yeah, so it's just nice having those people around that you can sort of bounce ideas off and around.

Aaron Phethean (29:30)
I definitely

like the sound of that. Some people are exploring the kind of much more structured mentoring, but there's a lot to be said for being in good company and being able to ask questions freely. And the kind of experience, senior people in the team are valuable. Joining as someone less experienced or coming in as a younger person in the team up your huge growth opportunity. Some really good ideas there.

David Draper (29:44)
you

Yeah, I I think that's,

I actually, so, uh, six years ago, um, I was like a, what they called a Google innovator. So people in education know what that is. So it's, you basically apply to support Google delivering their solutions and becoming an innovator. have to like create a video and share it, talk about your idea, what you might, how you can improve education. So what we ended up with, it was a lovely three day event, work in living, well, living in London, staying in London.

Aaron Phethean (30:21)
Yeah.

David Draper (30:22)
Um

for those three days and we're at the google offices And that's when I first learned about like design sprints and the crazy aids and they would actually bring in these experts to talk to you and it was like it wasn't quite a mentorship It was more like a guidance like that's They were you would ask questions. go. Okay. This is how I think you should do it You know, it's that kind of support and I really like that kind of element. It's not a

Aaron Phethean (30:37)
Mm, mm.

David Draper (30:44)
I do believe in bit of mental shit, but you you've got to go out and learn yourself as well. it's having someone to lean on, give you that bit of support, a bit of guidance. I think you learn by doing rather than having potentially a mentor showing you everything.

Aaron Phethean (30:59)
And I think perhaps traditionally they might have called that training, know, getting structured training at the job seems, you know, it seems a thing of the past. But, you know, there's so many opportunities to go out and attend things, learn things, find things out.

David Draper (31:04)
Yeah.

Aaron Phethean (31:17)
I would say most employers, universally, the employers that I've worked for and speak to customers, they encourage that. The time off to go and do that is no issue. It's hugely valuable. I do see some struggle to go out and find things or feel like they're almost allowed to or want to. I often think of meetups and things are in the evening, but I suppose if you want to

your career or your knowledge, then it's right there, the opportunity to just go and find it.

David Draper (31:49)
Yeah, there's, there's

a lot like I've seen, you know, there's events all the time, you know, personally, for me, obviously, got young children, it's I find it a challenge to sometimes go into into London, you know, it's not that far, but it's a challenge because I've got, you know, commitments. So I follow a lot of blogs, you know, get a lot of emails into my inbox. I try to keep up to date as much as I possibly can and

Aaron Phethean (32:03)
Yeah.

Yeah.

David Draper (32:12)
being involved in those kinds of discussions with other people, other professionals, you know, I'm still in contact with some people that did my data science degree with, and they're working in exciting places and having those discussions. sometimes those events, there's nothing beats those. you know, two weeks ago, was at AWS and met loads of people and having those conversations and actually, yeah, we could potentially do something like that. It's gathering those ideas. So it is good to get that out.

Aaron Phethean (32:27)
Yeah.

Yeah.

I'd love, I'd love this podcast to be one of those resources. And you know, we, we do newsletters, we obviously blog things as well. And I personally, I'm out looking for things, you know, often they're kind of challenge led or kind of an area of interest and you kind of read and go through them.

How are you going about finding them? mean, perhaps one of the things that we can share as a result of this is a list and collaborate on that, but that seems to be hard. How do you find things?

David Draper (33:07)
Yeah, don't, well to be fair, I've actually used AI to try and collect some resources recently. Like, you know, if I'm looking at a particular topic, I'll just pull in loads of resources. So Notebook LM, Google, you know, I've been using that quite a lot to try and create mind maps and things like that to learn certain things. Okay. Yeah. Like I said, email inbox, influxes from TLDR about AI. There seems to be about four or five emails a day.

Aaron Phethean (33:13)
and you're through.

So subscribe to them and that's sort

of generally filter through them.

David Draper (33:36)
Yeah, yeah, there's

lots of YouTube videos, if there's something specific, I would go and, you know, look at YouTube videos and, and think about how I might do it. There's quite a lot of people out there very quick and reactive to these things. You know, I think was the new chat, chat GPT last night or something released. And then I've seen about three or four videos pop up and watch this video. So I might stick it on. If there's a blog, what I found recently as well is

Aaron Phethean (33:57)
Yeah.

David Draper (34:02)
And if there's a blog post and I don't have time to read it, I will turn it into a podcast. And actually while I'm walking the dog at lunchtime, I'll put it on my headphones and listen to it. Yeah.

Aaron Phethean (34:10)
heard of that service, actually. Yeah, so it sort of summarizes

and then it actually, like you said, it turns it into more of a presentation.

David Draper (34:18)
Yeah, so it's like it, you know, if you were just reading an email, you know, that whole old robot robotic voice you sometimes get from, you know, Yeah, yeah, no, it actually turns it into a nice podcast. So I'm using again using Google for that. we just listened to it as a podcast on my so yeah, on my phone and my headphones walking the dog.

Aaron Phethean (34:38)
this is a slightly uncomfortable question. In our journey of becoming a senior in a role and experience often involves making mistakes sometimes. I wonder if there's an occasion that stands out to you that you might want to share or perhaps you don't want to share, but we'll share anyway.

David Draper (34:48)
Oof.

I wouldn't say a specific occasion. don't, I guess it was, I suppose early on in my career, you know, I about that risk list we were creating, you know, it was working on that. And, you know, we're trying to be reactive and it was like, it was pushing data to Google Sheets. And it was like, it was probably sharing it too quickly and being too reactive. you know, it wasn't.

we weren't in position to do traditional software engineering, which is QA and everything it was releasing it. But actually, that taught me that QA is really important and making sure that, you know, we were saying that this kid was absent when they weren't actually absent. And it took, you know, it took a few iterations, again, you know, it was a learning experience. And we were saying, we were very honest with the head teachers were saying, you know, this isn't 100 % accurate, we're working on this. But, you know, it's that evolution of

this is an iterative process. We're trying to get this right. We're trying to be really reactive, but you know, there wasn't a QA. It was just one person. And I think it's just leaning on one person. Yeah. To go too fast too quickly. Sometimes it's a bit of a challenge and you know, I like to, guess I like to do that. It's probably why I've been so interested in quantum like, yes, this is exciting. How can we do this? But you know, we're not ready yet and maybe I'm, we're not the right industry to do it. I don't know. There might be practical applications in the industry pulling.

Aaron Phethean (35:50)
Yeah, you can go too far.

Yeah.

David Draper (36:12)
We've got mass amounts of data from payroll, HR, accounting. might put it all together and create something fantastic in a few years time, but I don't think that use case is there.

Aaron Phethean (36:21)
So building a data team that's innovating and teams in general that are innovating, it's a challenge. we've talked about quantum and we've talked about things that are on the horizon, whether that's AI or quantum.

How would you advise people to prepare for quantum if it's to be the next game changing thing in data or in companies? And how would you advise people to prepare for change? Two questions there.

David Draper (36:48)
And to prepare for quantum, I think it's early, it's just keeping up with the industry, finding out what's going on. Microsoft have released a new chip. There's some negative thoughts on that and some positive thoughts, but I guess for the moment, I think it's just...

Aaron Phethean (37:01)
Yeah, it claims some

quite astounding numbers, but I guess the quantum industry is a little bit more...

David Draper (37:08)
Yeah, it's all about error

correction and making sure that there's no issues with error correction and making it easier, I guess. And it's, I guess, keeping an eye on it and seeing where it's going. I think a lot of people were astounded by AI and LLMs coming, but they've been around for a while. It's just, they were really pushed in the right direction at the right time, I think. And people have seen the benefits. And I think the same with quantum, will be, you know, keeping up with the times, you know, I think

Aaron Phethean (37:16)
Mm.

Yeah.

Exactly.

David Draper (37:35)
it might be people going, yeah, being aware of, know, actually, oh, no, our data could be leaked, you know, they could get through our encryption and security issues, that might be the first place to actually get people to think, oh, my god, this is, you know, this is a change, and we've got to prepare for it. But don't know, you know, I would love for it to be really quick and impact the industry. But actually scary.

Aaron Phethean (37:36)
Take it away.

Yeah.

But as with everything,

it's quite hard to forecast. It's not like with AI and, you know...

David Draper (38:02)
Yeah.

Aaron Phethean (38:06)
Initially, there was an awful lot of excitement around LMs and generative transformers. The fundamentals, though, remained the same. I'm obviously in a data company where we're helping our customers gather and process data. That's sort of a fundamental of AI, that you will still need that.

David Draper (38:25)
Yeah.

Aaron Phethean (38:25)
you still need like, you know, you won't have knowledge if you don't have the data. I wonder if with quantum, there's also a kind of, you will still need, you know, there's a sort of a root investment that would still be worth making even before quantum is, you know, arrived.

David Draper (38:46)
Yeah, I think there's, you know, there's going to be a need, know, storage is going to be an issue. You know, I know they're already looking at how to store data like DNA and some interesting things around that.

Aaron Phethean (38:56)
So because what's

special about storage and quantum - the link there?

David Draper (39:00)
I think it's, you know, there's a lot of data that you're going to require to process. So they're only going to mainly solve MP problems and challenging problems and DNA sequencing things that need a lot of data throughput. And I think that's where do we have enough data to do that? And I think that's still at the moment, internally in a lot of businesses, do we have enough data to do machine learning? And the probably answer in quite a lot of businesses is no, and it's probably the same for quantum, you know, are we in a position to...

Aaron Phethean (39:12)
Right.

Mm.

David Draper (39:28)
have enough data do we, you know, to make those predictions. Again, it's it.

Aaron Phethean (39:31)
Yeah, you mentioned

earlier the kind of multiple paths and exploring multiple paths. I mean, that seems like maybe a structuring of how you go about, determining the, the options, there's algorithmic approaches that are still worth investing in and, perhaps more conceptual than, quantum changes, some of the ways those algorithms work.

but perhaps an investment in algorithms is still worth the time and effort.

David Draper (39:59)
Yeah, I think what a lot of the machine learning sort of quantum ML is utilizing both. it's like leaning on traditional machine learning, but employing quantum computing circuits to improve the efficiency and speed of how it comes to that result. So that's what's quite interesting. so, know, TensorFlow, Cirque, you know, they're working together to develop a coexisting model, I guess.

So I think all the industries talk about quantum LLMs, it's that kind of thought. They're just adding quantum in front of it because it's utilizing quantum power. it's, I guess it's traditionally moving away from, we were building algorithms on a classical processor, CPU unit. Now we're doing it on GPUs and Nvidia released a new GPU last week and a desktop. Then if you saw the nice little cute little desktop computer or GPU unit they've got.

you know, that's where it's going at the moment is it's the processing power.

Aaron Phethean (40:56)
So that was designed

in a completely self-contained LLM, wasn't it? Like it was designed to run almost like isolated LLMs, isolated assistance, general idea.

David Draper (41:02)
Yeah.

Yeah, I think I we've probably all

looked at the most powerful computer we can have on our desk. It's one of those isn't it? I it's cost 50,000 pounds or something and it's got four GPUs in it. But you know, I'm not going to need that. you know, might have someone sitting in, sitting in research science with lots of data that they can't put on the, you know, put on the third party internet and cloud infrastructure, they need to do it locally. That's quite interesting.

Aaron Phethean (41:14)
You

Yeah, yeah.

Yeah.

I mean, way

of technology and we will demand more. will do more. You know, this is, this is sort of one of the fascinating things about, being a vendor of technology. we see that the more we lower the costs, the more, well, you see demand, the more use it gets. So, you know, it can be quite some quite funny situations, the rising cost of cloud or snowflake or data warehouses in general.

ETL platforms, they then get used more or that might be rising costs because you want to acquire more data. There's something driving that rising cost. Other times there are perhaps a strange pricing models emerge, actually more demand will generally lead to more cost, but then a lowering of costs as you start to react to that and build technologies around it.

David Draper (42:07)
Yeah.

Yeah, I think I've kind of seen it in electric

cars, know, BYD last week released the new battery, it be charged in five minutes. And that's where quantum has been, you utilized in the research in improving battery efficiency. And I think that's a really interesting piece, you know. But again, that's a whole other interesting industry about hydrogen cells and all that sort of stuff. But it's how I see it.

Aaron Phethean (42:42)
But why not, why not have a look around and see what else is going

on in other industries. So it's great advice.

David Draper (42:46)
It's a facility.

It's like, you know, I see, you know, computing algorithms is all a facilitator to improve different industries, you know, research into different industries. I've always been interested in that kind of science background. My, my, my dad was in science and working for Pfizer's and, you know, it's always been a fascinating area, what they've been developing. And it, it's that kind of approach, quantum computing, AI should be able to help improve people's lives, make them more efficient.

still think the West is still so far behind in terms of, you know, China shoving loads of money in this and I don't understand why we're not.

Aaron Phethean (43:23)
Yeah, so

there is a, I mean, there is a bit of a, it's a strange sort of reluctance to like engage with it in a really significant way. I mean, the UK has talked about a couple of things.

David Draper (43:36)
But then

we're behind and in education, in education, they've actually in China, I saw a post the other day saying that all students in China have to have at least eight hours of learning AI in the curriculum every year. And I said, what's the UK and US and Europe, whatever? Are they doing that? I don't think so.

Aaron Phethean (43:38)
Yeah.

Yeah.

think that, I

also look at the UK and I think, you know, we were...

So we, mean, obviously Kiwi and not necessarily from these shores originally, but I look at what they're doing and investing in, and it's almost like that research, science, education, that's like, have been under invested for decades. And they wonder like, why have we lost our supremacy in technology? Well, it kind of looks obvious.

David Draper (44:03)
Hahaha

Yeah, I just...

If you think, you know, all the research scientists that you you you're in it for the love of it, you're not in it for the money. Whereas if you want to earn money, you would move out into business into the real world, I guess. You know, that's, that's how it's perceived. Yeah, yeah. Or, you know, or it's, you know, Google did that 20 % time. And I didn't know of any businesses doing that 20 % time. I would actually, you know, people talk about the full day working week, I think.

Aaron Phethean (44:39)
Maybe it swung too far that way and this has not been attractive enough.

David Draper (44:54)
I would love to four day working week, but that one day that you're off is learning time. You're doing something or there's an offering out there that can deliver a nice interesting area. And again, then you support the development of a product or development of the business because you're learning new innovative ideas. And I think we've lost that.

Aaron Phethean (45:11)
Yeah, I definitely see

that. That's like you see a lot of people talking about the four day working week as a benefit as a little bit just work less. I mean, it doesn't feel like work if you really enjoy it. So if your fifth day is like something you really love and invest in time and energy into might be really closely related to your work, but the time and the investment there will be more than paid off, you know, discoveries and the learning.

David Draper (45:37)
Yeah, like,

yeah, like, obviously, I come from a teaching background and you work when you get home. So after you put the kids down, you work in the evening to mark and whatever. And that's the frustrating piece. I was like, I want to learn. don't want to, you know, I don't want to always be sitting there planning a lesson. I want to learn and I kind of do that now. My wife's a teacher, she's planning lessons. So I go and learn and tinker around why she's doing that. So,

Aaron Phethean (46:00)
So one question do you wish people would ask more about data strategy or your work?

We did talk about where can listeners connect with you, learn more about your work. Do you do external? Do you produce material and educate others? you write externally? What do you do?

David Draper (46:17)
I would love to.

I'm going to be caught in it. I feel like I don't have time. know, the day job and everything. I don't have time. I'd love to write blogs, but it just swings and round about it because I've got people who have actually come and given me their CV and they've written all these blogs and they're writing a blog a day. And I'm like, I don't know how you have time. I generally don't have time to write a blog a day. You know, I would love to do more of that. you know, coming from a teaching background, but

I feel because of my team, I'm supporting them and supporting the business in teaching them. have these architectural reviews and we bring in architectural reviews. We've brought a recent data mesh design to the architectural review and thought about how we change the way we process data to make it more flexible. It's kind of educating people on what that is. They're not everyone's experts in your field, so it's trying to explain that in a simplistic way.

Aaron Phethean (47:01)
Yeah. Yeah.

David Draper (47:11)
way for them to understand quickly. You know, that might be more complex than what you're trying to explain, but they don't need to know the complexities. And if they're interested, then they can come and speak to you. And that's where I feel like I am open. I'm happy to, you know, there's no

Aaron Phethean (47:23)
Yeah, more internal

communication and more internal focus. Definitely see the benefit of that. Doing it externally, I sometimes failed to see the link between if they just enjoy it or actually what they're trying to achieve with the kind of external content.

David Draper (47:40)
It's the same as you've got a work problem, you need to solve a work problem, but I'd love to write about it, but it's work data. Sometimes those problems we don't want to share, it's just one of those things. It might be a unique problem I've solved, but I can't write a blog about it.

Aaron Phethean (47:51)
Yeah.

Some companies

are much more open to that. know, obviously Netflix have released game changing technologies, research papers. Not all companies are comfortable with that. I wonder if you have any advice or any experience seeking that approval. Because I come across that even with these podcasts. They don't necessarily have permission or feel comfortable and that might even limit them coming on the show and talking about it.

David Draper (48:02)
Yeah.

But I think, I think that's the problem. think, you know, I saw your recent, you know, people talk about not using AI in the workplace and people banning it. like, why? Like, you know, everyone's got similar problems. I'm not talking about specific clients or specific issues we've got. It's about solving a common problem. And if we can do that as a community, you know, and that's what I love about Netflix, you know, ice-spaired tables, it solves a problem.

Aaron Phethean (48:32)
Yeah.

David Draper (48:52)
and they've built that technology for everyone else to use and it's sped up development. If people didn't do that, we wouldn't be where we are.

Aaron Phethean (48:57)
I think you're

educated background, you're leaning towards open source. This is sort of a definite mindset that not everyone shares that if we all benefit from it, we're all going to be better off. So why be protective in nature?

David Draper (49:16)
Yeah, it's probably my, yeah, like I said, a bit fluffy. I, you know, I want to tech for good. It's that kind of feeling about it. And I understand there's intellectual property and things like that. It's, but if you're not sharing the ins and outs, it's more sharing the technology that helps you do that. I think, yeah, it's maybe it is my background, but I don't see a problem with that. You know, Netflix have released probably what they could have sold for millions really, and what they've done and they've open sourced it and

Aaron Phethean (49:43)
Yeah.

And everyone benefits. And I actually, feel very much the same about what we're doing with Matatika We kind of have this idea of an open source core, no vendor lock in, being able to run it and outgrow us is like sort of our expectation. That's like almost an unnatural approach for a software vendor. That's not protectionist at all. That's everyone's going to benefit if there's a better core technology available.

David Draper (49:46)
And then, yeah.

Aaron Phethean (50:08)
And yeah, I've definitely aligned with that myself. I have found in the open source world that can be like, Oh, but you're not really doing open source properly.

You know, and like sharing ideas is as valuable as sharing the code or making a library. And there's some pretty big negatives of, you know, individuals having to support libraries that are really infrastructure and virtually no way to fund them.

David Draper (50:20)
Mm.

Yeah, yeah, I get that

as well. It's, it's always two sides of the coin isn't there, there's always that, you know, this is, this is a great idea, but who's going to fund it, who's going to pay for it, who's going to pay for people's time. And it's, and you're relying on people that are interested in it. And you know, I'm on the Iceberg emailing list, and I'm impressed by the amount of people that reply to it. So, can we do this? Can we do that? I've got the code here. And it's like, wow, like,

Yeah, that's amazing. I wish I had time to do that. I wish I could stick my head in and do that, but I don't. And I guess it's pat on the back to them who do that and love it and involved in that.

Aaron Phethean (51:08)
Yeah.

Yeah. And why not scratch that itch? Well, thanks though. Thanks again for sharing and I look forward to hearing more about the quantum supremacy.

David Draper (51:21)
No.

Yeah, we'll see.

Aaron Phethean (51:25)
And perhaps a good point to wrap up on that, you know, explore the technology. Yes, you explore the value of the technology and, and, know, think you try and understand how to get benefit from it, rather than fear it. And so, yeah, thank you, David. That's been a fascinating delve into data and your career and, and also quantum technology and how to leverage it. So thanks.

very much for coming on the show.

David Draper (51:51)
Thank you. Appreciate you asking me.