Interview with Martin O’Leary
24 May 2021
John Howe, Manager with Brightwater’s IT team is speaking with Martin O’Leary, a Data Scientist who he recently placed in a role.
JH: Hi Martin, thanks for joining me today. Thanks for giving me your time today to have a short chat about the market and also your experience as a job seeker in the market as well. Thank you for that. So I suppose we’ll start with what we do. Can you give me a quick overview of what you do as a Data Scientist? What do you as a Data Scientist actually do?
MOL: I don’t think there’s a quick overview.
JH: That’s a fair answer!
MOL: As a data scientist, I’m generally the one responsible for all sorts of analytical activities, mostly project based where I’m coming from, so I’m not embedded in a business with lots of domain knowledge. I’m more of a project-based data scientist, going from different parts of the business and helping them realise value from their data, their latent value as we call it. That involves all sorts of stuff like data exploration, taking anything from Excel sheets to images to sound and trying to wrangle that into something more structured, like numbers, 1’s and 0’s and all this kind of stuff, then exploring that, trying to visualise that and seeing if there’s patterns before going down into modelling at this point, it’s more like exploratory stuff – is there a hypothesis that we can put forward to our business, to our clients and say “by the way, we think that you have this problem you’re telling us about, we can see something in the data some way that we can help with that”.
At this point in the project, and I guess as part of the role, there’s a lot of dealing with uncertainty in terms in where the project could go to. The way we try to deal with that and reduce the risk there is that we really try and get to know what the business wants, what their key objectives are and really talk to them about that and make sure we understand the business context. As I was saying, we’re in a kind of central role a lot of the time as data scientists in that we’re not embedded in the business, we move from different domains. When you do that, you come into every project with a little bit of knowledge about the business area or the domain but you generally need to rely on subject matter experts for that so there’s a lot of dealing with subject matter experts, trying to upskill and get more knowledge about the domain.
Later on, what you do a lot of, is data processing. The calories you expend as a data scientist is in data processing. To do that, to make the best use of the data you have, it’s better to have that domain knowledge and you need to work with subject matter experts to get that domain knowledge.
That’s a lot of the upfront stuff, then the middle, the boring stuff which is the modelling and thinking about your hypothesis and when that makes sense to the business stakeholders or the subject matter experts, “we agree with that, that goes along with our key objectives, it looks like you’ve got a pattern that you want to go after and a hypothesis so let’s all go down that route”. Then there’s a lot of hands-on data processing, coding basically that’s a bit unstructured at this point, you don’t really make a lot of scalable things at this part of the process. We work in an agile way where you get quick feedback from the subject matter experts and the stakeholders in the business every two weeks. Then at the end of that, you’re moving more into DevOps, ML Ops delivery part of it so then you put on your engineering hat. Honestly, you have your engineering hat on all the time but at this point in the project from the outside looking in, the engineering part starts when you put stuff into production and deliver it or make it useful. Because a model on a Jupiter notebook really isn’t useful so something that you can share with stakeholders or even get into production quickly, you need engineering involved there.
JH – Really interesting.
JH: So I think a lot of people would be familiar with or not so familiar with what a data scientist does, they see the word “data” and they see the word “scientist” they think “right, we have a traditional stereotypical nerd who sits in the basement and types away with 1’s and 0’s and Excel sheets when it’s purely research. What’s interesting to me was when I first started working with data scientists and data analysts all those years ago was that you need really strong communication skills. Ultimately you’re problem solving aren’t you? In order to be able to understand how to be able to solve a problem you need to ask questions that will work for you to enable you to find a solution. That always surprised me at the beginning but it’s very interesting. What other skills does a data scientist need?
MOL: Yes and I didn’t expand on that I didn’t even go into that side of it. A lot of what goes on in data science is smoke and mirrors to the outside world. To be able to build confidence in what you’re doing, you need to be able to explain what’s going on there. A lot of people are taking rule-based, “here’s things and let’s try it out, AI, machine learning, deep learning etc” which they don’t understand but they really do understand business rules and they’re going from that to they’re trusting us to build this model that they have no insight into so “explainability” is huge there. You have to be able to communicate that to people and explain what a model is doing from simple cases such as logistic regression or linear regression all the way up to ensemble methods and even deep learning. Nothing is going to work, nothing is going to make it all the way to getting business value unless you can build confidence and bring the people along with you who are going to make those decisions to put that into production.
JH: Yes, that are very good points and I’m seeing a lot more of that as requirements from my clients, from employers who are looking for those characteristics. Historically they would be so focused on the tech experience, the tech stacks that candidates would have worked with, that’s still important but less important now as more and more people have exposure to those tech stacks and those technologies. Now they’re starting to look at those softer skills in data science and across data analysis traditionally and across innovation engineering as well. It’s interesting to see the requirements around what a data scientist is and what’s expected of a data scientist from future potential employers and how they change over the time. That’s been my experience and glad to hear it’s been your experience too. We’re on the same page.
JH: You’ve touched on a few thing there, deep learning for example, AI which you can’t escape, we’re all thinking about AI at the moment, it’s endemic and it will continue to be so. Even in my own industry, artificial intelligence is something that’s a part of what I do in terms of the search engines and the databases that I use, are driven by artificial intelligence and machine learning as well. That’s been interesting to see it eke its way into my job which has traditionally been quite phone, email, (not Roladex, I’m not that old but it has existed at some point as well). So what I’d ask you regarding trends in the industry that you’ve seen, I’d say we all probably know the trends but where do you see the future of your industry or sector taking you? Where do you see the future of data science?
MOL: Honestly at the moment, you see most of the roles being advertised from a data scientist’s point of view in startups and tech forward and everyone who’s all about using data as part of their business model as opposed to a side effect of their business. Activities, an offshoot, let’s say insurance companies, their general business model doesn’t revolve around AI but a lot of data is included in their data-warehouses and they benefit from it. At the moment, they’re not the big hiring people but eventually I would say that these big slow moving ships, they’ll take a bit of time to move in the direction where they’re going but when they do, I’d say you’ll have all these big traditional companies, (you might look at them and say they’re not agile but eventually they will have to be) and basically you’ll see data science as just a normal function in every company. I think, because nearly every company in every business that you can think of, there’s data being generated at a crazy rate at the moment, and every business can benefit from that. They can unearth intelligence or value patterns, there’s different use cases but I think to be concise about it, there’s a lot of hiring right now and there’s a lot of availability in the tech industry around data science and I think, as it gets more mature and as bigger companies get less traditional data centric companies, get more value. I don’t want to be using terminology and make them sound bad but when they do change their view on these things, it will become a more regular function like HR or finance in a company.
JH: That’s cool and makes sense. By the way, slow moving ships, I’m going to steal that. That’s well said, it’s very descriptive and I know exactly what you mean by that.
JH: Another question that I have would be about your experience. You’ve recently moved jobs and congratulations again. What I wanted to ask you was a couple of questions I wouldn’t be much of a recruiter if I didn’t ask you about your recruitment journey. When you started your job search, before you looked out there into the market, I feel that some candidates might be thinking “right we’re still in a pandemic (we’re coming out of it thankfully) but we’re still in a pandemic needless to say and that’s going to have an impact of jobs out there. Maybe I should stay where I am and see how things change for me”. I suppose is that something you were thinking about before you started looking? Why did you decide to change roles, particularly during a pandemic?
MOL: Yes, there’s always the two guys on either shoulder, one is saying “stability is good, stay where you are” and stuff like that. A smooth sea never made a skilled sailor.
JH: Another nautical related reference. I have to give it to you Martin, you must be a sailor, are you?
MOL: I can’t even swim. Most sailors can’t anyway.
JH: That’s true.
MOL : I guess what I’m saying is that it’s easy to sit around and the inertia gets much bigger then to get you to move. So there’s the whole pandemic, uncertainty, there’s a lot of anxiety in social terms about change and everything. Funnily enough though, for my role and roles I was looking for, now is quite a good time because the nature of my work and where I can do it and the work-life balance, at the moment, companies are changing a lot in how they look at that. Most of the roles that I looked at and I saw were available are all remote first. That’s something that’s huge to me because I have a small one year old and another one on the way so for me, the work-life balance (I know this is a side thing now and it’s the same for everyone and I’m lucky in that I can avail of working from home) but I really benefitted personally at home from not commuting for 10 hours minimum a week. A lot of the roles that I saw out there were remote first so that appealed to me. I guess from the personal side of things, there’s never a right time, you can be waiting forever for the right time to move. There were a lot of opportunities out there that excited me.
JH: Yes, absolutely. That’s what we see a lot of now. We have never been busier. Recruitment in general has never been busier in years. There’s not enough time in the day. I feel like those 10 hours of the week that I’ve been saving with commuting into Merrion Square each day is now gone from covering all the jobs that I’m working on at the moment. It’s a great complaint to have in my line of work, and increasingly, you’re spot on the money there, Martin, in that all of our clients, most of them have adopted “remote first” and will continue to do so. I had a call yesterday with a relatively new client, a very large consultancy on the data side. The director of the business said that when the pandemic is done, people will still continue to work from home as they have been. They’ve proven that it works and what’s the point of changing something if it works. He’s dead right and I think a lot of companies will look at that. Companies at least for now are only starting to look at those policies. We see a little light at the end of the tunnel and as the timeframe for true “opening” and true “going back to the office” becomes clearer and clearer, companies are starting to look at their policies. It’s only when they sit down and discuss them properly in a round table discussion with their stakeholders that they will be able to say “maybe we shouldn’t change it or give it as an option of working in the office” because lots of people like to be working in the office including myself. I think that’s going to be a big thing and decision maker for candidates when looking for a role.
MOL: Yes. For me, yes.
JH: The actual interview process for yourself, say for example when you first had your initial chat with myself and your experience after that, would you be able to describe that please? What was the job hunting process like and why did you use a recruiter like Brightwater to help?
MOL: When I first had a chat with you, the first thing that struck me, I’ll be honest, from my point of view, from someone looking for a job, a lot of times, you can get the feeling that you’re one of a hundred people that the recruiter is going to talk to that day and you’re not really, it’s a very quick brief, an impersonal thing. But the first time I was talking to you actually, there was no kind of….., it was more about my needs and stuff, it wasn’t about a particular position which was refreshing. You don’t get that all the time.
I found our first conversation was very good, you kind of were very eager to make sure you found out exactly what I was looking for because, I suppose from your point of view, you don’t want to waste your clients and your own and my time looking at jobs I was never going to be interested in. So that was good and that wasn’t something, (I had dealt with different recruiters before) and that wasn’t something I had experienced so that stood out for me. Your knowledge of…. I only see what I see and I probably saw my interpretation of things, my understanding of the market would have been what I saw on LinkedIn so you were able to give much more, you were obviously much more knowledgeable about it and that comes across too, you know? I would say, having gone through the process and yes, I’m lucky to have picked up a job, I don’t think you’re very likely to be successful if you just kind of cold call these companies yourself. I don’t find that works, the scattergun approach, well, I didn’t have success with it.
JH: Yes, I think that you’re spot on. From a recruiter’s perspective, number one at the moment, most recruiters operating in data science or in the tech space in general wouldn’t have time for cold-calling companies anyway. It’s quite time consuming. It is a fundamental part of any sales job, I feel and that’s unavoidable. Recruitment is cyclable and there’s times when I’ll have plenty of time for it and that’s just the way it is. But I think you’re right, it’s about leveraging networks, building on those networks and even with a strong network, you still may not have success in terms of opening up jobs and that can be down to many different reasons. Time; people just don’t have time to engage as well, these are the challenges of recruitment. But listen, ultimately you were successful so congratulations as well.
JH: Just your experience with remote interviewing. I’m quite curious about this. Only yesterday I’ve had somebody specifically say “I wonder if the hiring manager is open for a 20 minute coffee” and I asked “do you mean in person or having coffee over Teams or Zoom?” and they said in person. It’s not the first time I’ve been asked this and it’s a funny question. I’d say there’s quite a few hiring managers, especially at the more senior levels, more strategic hires who’d be open to it potentially and candidates as well. So what I know from that, and I know from people asking these questions and we can only assume that the people asking that it’s only the tip of the iceberg, that there are people who do want to and just don’t ask because pandemic etc, I wonder how did you find the remote interviewing experience yourself as opposed to in-person interviewing? Remote interviewing is prevalent now but how did you find the process?
MOL: I’ve a few thoughts actually that came into my mind while you were talking there. It can be very impersonal by video etc. We’ve all seen plenty of little jokey videos of people talking over each other on Zoom calls.
JH: The cat filter ones as well Martin, I’m sure you’ve seen that. That’s one of my personal favourites of 2021 so far.
MOL: Yes, it’s one of the best things I’ve seen but those kind of funny things aside, I find that, let’s say, the barrier to interviewing people is lowered a little bit. This is just my view, I found that I had probably five, maybe six interviews for different roles and I think that for a few of them, I felt like I was in a long list of people. That day was probably booked out 8 hours for 30 minutes each so I think from the company’s point of view, it was good, they can get through interviews quicker as opposed to calling people in. It’s easier I guess to arrange times, I think personally. The other side of it, the side that’s going to be a challenge is that I’m joining a new company now and I don’t know when I’ll ever meet my co-workers face to face for the first time. I did specify when I was looking for a role that I was looking for remote first, but I feel that for the first six months at least in the job, it’s good for a couple of times to see your co-workers and meet them face to face.
Company culture, it’s a big thing to embed yourself in your company culture and get up to speed really quickly with the ways of working as opposed to being remote. Different companies will be better or worse in how they on-board people but ultimately you’re trying to embed yourself in a new culture and all businesses will have different cultures. My hiring manager was based in New York so the coffee, if they would have paid for that, that would have been a brilliant coffee, a 20 minute coffee for me to go over there. I do know what they mean. A phone call, you lose a lot of body language. A Zoom call, you’ll probably get a bit back. There’s a lot lost as well not being face to face and in person. For me, I didn’t find that too challenging, that was just my experience but different people are different. The roles that they’re looking for, I think that the different roles and the different profiles of people will feed into that very differently. I’m very much on the engineering side of it, the kind of writing code side of things so for me the remote bit is good and it’s fine but I do foresee, not an issue per se but it will be different trying to embed and feel like part of the company. Nothing has changed, I’m sitting at home and I’m gone from working with one company last month to a different company this month but I’m still sitting at home. My day to day, my life hasn’t changed in a drastic way. When I moved from my previous company to the last company I was working with, I went from driving over here or cycling over here or getting two buses over here, my whole life changed, everything changed.
JH: Good point, that’s actually a really good insight and even for me as a recruiter, that’s a good point for me to think about.
JH: My penultimate question is, so your new job, you’ve spoken a little bit about it and again, congrats. But what does this new role then mean for you in terms of your career, your trajectory, and then you as a person as well, if you wouldn’t mind sharing that with me, I’d be very curious. What does your new role mean for you, particularly in terms of career development?
MOL: So, I might have given a bit of a hint as to some of the frustrations I was feeling in my old place there, the slow-moving ships and all. I was working in, let’s say, a classic insurance company and I was kind of in a cross-functional team and working across different lines of the business which also makes things nice. You have a very varied workload and use cases are all different and you’ve a lot of variety which is nice. On the other side of things, you’re kind of in and out quickly and ownership isn’t necessarily something you need to worry about too much. You don’t see the business value all of the time. It’s almost like a consultancy which is how I was working most of the time. So I, previous to my last role, was very much a domain expert or a subject matter expert, whatever way you want to call it, I was embedded, I knew the business really well. I liked that, I liked being knowledgeable in the business. This new role is very much more in that direction. First of all, it’s a data scientist role so end to end ownership is expected from conception all the way up to delivering and maintaining modelling and production so that’s something I really want to do. My undergrad was mechanical engineering and I did my Masters in Data Analytics but I’ve an engineer’s brain and the engineering side of things is what makes me tick. The ownership, the maintained ownership and being embedded in the business is important. The business itself is data, centred around data. They’re not quite a start-up, they’re 8 years old so they’re past the early start-up phase now and they have the proper benefits and everything else.
JH: They’ve gone past the scaling stage at this point.
MOL: I guess so. Even though they’re still small in numbers, they’re branching out into new businesses and other stuff.
JH: Well, that’s good.
MOL: Yes, you’re not……, the thrill of working with a start-up isn’t one of the benefits they’d list. You get health insurance instead.
JH: I think I’d prefer the latter myself, considering where I am in my personal life and with the threat of pandemics and what not, I think it’s probably more important.
MOL: Exactly so that was something that put me off going near start-ups a lot actually, was I suppose where I am in my life. There’s a lot of room to grow where I’m going and there’s a clear path of growth. There’s a function there of data science and there’s levels and there’s growth and everything else. Where I was, was very SEAL Team Six, cross functional team, plopped into or embedded in this department but we were not necessarily really fitting into so the growth path, if you wanted to grow there, it was all managerial to fit in with the rest of the business and not necessarily data science or tech side of things if that makes sense. So the growth path was not aligned with my career goals there either. So (in my new role) there’s growth, there’s company culture which is a big thing.
JH: Yes, I think company culture is super important, even more so now where communication and collaboration skills are paramount. I was talking to one of my colleagues this morning and he was saying that to get something done, you could just shout across the room to somebody to get an answer. Now you have to call around the houses, it takes ten times longer so there’s a frustration there as well and I think a collaborative community, so to speak, can really help a company survive or not do so well.
MOL: That’s a tough one, always a tough one. In a previous life, I was a developer, developing applications and I was sitting luckily near my end user. I could develop something, go quickly to them and ask them what they thought of it, “do you think this works for you, suits your needs etc”. It was super hyper agile in terms of feedback loop. I struggled a little bit as well in the last place, in that we had no standard messaging solution so we were trying to figure out the best way of keeping in contact. Meetings are the bane of a lot of people’s existence and pre-scheduled meetings for the sake of it, they just turn into “so anything else?” and it’s just filling in time. The new place, the culture is very much they have just one recurring meeting per week and everything else is synchronous. That’s something that really appeals to me as well. When you lose that and are not able to walk over to someone, ask a quick question and get a quick answer, using technology to try and replace that can be tricky.
JH: Yes, absolutely agreed. It’s been over a year now that people have been working from home and I still think that whilst the technology support is there and it will continue to change and develop and evolve, I think that you can’t really replace that face to face human interaction and communication. We have literally millions of years of evolution aimed at making that the best thing for us.
JH: I suppose the final question which is a short answer I presume, and hopefully in the positive which is, based on your experience working with me, working with Brightwater as a candidate and that entire experience and process, would you be inclined to use Brightwater as a recruitment partner if you were a hiring manager? Given your experience as a job seeker with Brightwater, would you use us again if hiring?
MOL: Oh yes, I would. Just knowing from the other side that the communication lines were open nearly always. I was never left wanting for an update or anything like that. Having been through that side of it, there’s no doubt that if I was hiring for a role, that you guys would be in contact with the right people and you probably would do a good job in filtering out people who, for a variety of reasons, wouldn’t be a good fit in terms of what they were in for. So definitely I would.
JH: I appreciate that, thanks so much. Thank you for your time today, Martin, I do appreciate it. I’ll leave you to get back to work and also to think of more nautical themed phrases which I have to say were very amusing and I’m definitely going to steal one or two of them. So thanks again, take care, have a good one.
MOL: Cheers, bye!
John Howe, Manager with Brightwater’s IT team in conversation with Martin O’Leary, a Data Scientist who he recently placed in a role.