Mark Douthwaite is an AI engineer at Peak, a start-up which works within retail and ecommerce, adapting AI to solve supply chain issues with the likes of ASOS, Footasylum, Fred Perry and Morrisons.
So, let’s hear what Douthwaite does every day, and see if he can cut through the hype and explain the true power of AI.
Describe your job. What do you do?
I’m an AI Engineer, which is a role that sits directly between engineering and data science. I utilise my knowledge across both of these disciplines to help bridge the gap between the two, and convert data into actionable insights that can power growth and increase efficiencies across a business.
Where do you sit within the organisation?
I officially report into our head of data science research and development, which then feeds into data science and engineering. I really feel that this is a great idea from Peak as the role brings together two often siloed areas (engineering and data science) and enables the two disciplines to work together to make the most of out each other. We’re seeing a lot of other tech firms doing the same. However, you do have to have some soft skills to deal with this challenge.
What kind of skills do you need to be effective in your role?
You need quite a lot of computer science know-how and software engineering experience to figure out how to actually turn interesting data science ideas and insights into something that businesses can really rely on, day-in and day-out.
In addition, the soft skills needed in a role like mine are all about being able to juggle the priorities of many different things at once. It’s about understanding those priorities and balancing them when communicating across the different teams. For example, our data science team works to deliver fast, reactive turnarounds for our customers. So, when we get a request, we can start preparing prototypes and ideas to go back to the customer pretty quickly. However, a lot of the engineering tasks can be slower-paced due to the nature of an engineering development lifecycle and all the associated stability and performance considerations that come with it. This means that the priorities of the two teams can be quite different. It’s all about establishing clear lines of communication across the two to quickly resolve issues and get feedback on our products as fast as possible.
Tell us about a typical working day…
I think the only thing that’s typical for me are my mornings! Which is to get in and then tackle the administration, emails – all that good stuff. Then I do some code reviews, if there are any up for review. After that I might dig into a bunch of productivity tools, just to see what the day looks like and what needs doing as a priority. From there on, my day can vary dramatically.
What do you love about your job? What sucks?
We’re a fast-growing company and every day is genuinely exciting and different, which is what attracted me to Peak. Due to our size, we get to build and design things and explore new technologies in a way that maybe wouldn’t be possible in other organisations that have more embedded processes, where it’s harder to introduce something new or different.
As we’re also an agile team; I get to do the types of work that I might not be able to carry out elsewhere. For example, I also do a lot of proof of concept projects and software design work, so that keeps the role really interesting. People here are very open to new technologies and ideas, which is great for someone like me who likes to build things.
What are your favourite tools to help you get the job done?
I use Trello quite a lot, and the engineering team uses Jira. We use JupyterLab and R Studio as data science environments. The development environments that I often use are VS Code, PyCharm and other IntelliJ IDEA tools. At Peak, we also have our own proprietary tools and platforms. The main one, naturally, is our own AI System. It allows us to handle huge amounts of data at scale and speed, and enables businesses to productionise AI, end-to-end, in a single platform.
How did you end up at Peak? And where might you go from here?
I was speaking with some fellow engineers that work for a well-known software company, and I was asking them what they would do if they started their careers all over again. They said they would look for a fast-moving, young company full of interesting people doing interesting things so that I could be exposed to all the different sides of the business. So, that’s what I did!
It’s important to really feel what it’s like to be able to make your own decisions and lead on projects, which is far easier to achieve with an agile, growing company. So, when I saw Peak, I felt that the team here was doing some amazing work, and is probably one of the most interesting commercially-focussed AI companies in the UK. In the future, I’ll stick to the same strategy – for me, it’s all about making technological contributions.
Which marketing or ecommerce has impressed you lately?
I’m particularly proud of our work with Footasylum, a UK-based sportswear retailer, which wanted to invest in hyper-personalised marketing communications. We worked to analyse its customer data so it could have a view of the people most likely to be both engaged with the brand and in the market for its products.
From these customer profiles we created bespoke algorithms that distributed hyper-personalised product recommendations. Our work achieved a 28% uplift in revenue per email sent by Footasylum, and when we implemented the technology into its social media advertising campaigns. In another solution for Footasylum, we achieved a massive 10x higher than industry average return on ad spend (ROAS) across the campaign.
Do you have any advice for budding data scientists that want to work in retail?
One of the seemingly unifying things about data scientists is that they’re usually very curious and very persistent. It’s important to foster this mindset and always keep learning. This field can move at such a quick pace, and you can get left behind. If you try and foster some core skills in mathematics and statistics, with a bit of computer science on the side, it should make it easier to stay up to date. It’s just a case of always wanting to learn more, continuing to grow and generally staying engaged. There are so many new things coming out in data science that it can sometimes feel overwhelming, but you just have to roll with it. Don’t worry too much!
Comments