Meet the Team at GoustoTech — Marc Jansen — Data Scientist
This post is the fifth in our series interviewing members of our GoustoTech team. In this post we meet Marc Jansen who has been part of our Data Science (Artichokes) team for the last 18 months.
Please introduce yourself and tell us about your role at Gousto.
I’m Marc Jansen, data scientist. My role at Gousto involves developing and deploying optimisation and machine learning algorithms. As a team we look for opportunities to optimise and automate various business processes (e.g. warehouse optimisation) and enable new service features (e.g. personalisation) and build data products to leverage these opportunities.
How long have you been at Gousto? How many boxes a week were we selling when you joined?
I started on January 3rd 2017, so just over a year now. I had to look this up: I think we were doing about 11k orders per week.
Can you tell us a little about your journey until arriving at Gousto?
Before starting at Gousto, I completed a PhD in Management Science. In fact, I was still working on corrections well into the first few months of working at Gousto (don’t worry, my team were aware). My research was on applications of game theory to structure incentives in technology service contracts. Though a far cry from algorithms and machine learning, the PhD gave me a lot of exposure to constrained optimisation and probability theory. To my surprise, I use this knowledge very frequently in my role.
What made you decide to join Gousto?
As part of my PhD programme, I did a lot of teaching. Through teaching, I met Timo (our founder) about a year after he’d started Gousto. His stories about Gousto’s growth and where he saw things going sounded really exciting. Then, as I was gearing up to finish my PhD, Timo reached out again and asked me to consider coming in for interviews. Naturally this was an offer I could not refuse…
What has been the most exciting thing you’ve worked on at Gousto since joining?
One of my first projects was building an automated stock control solution. With both subscriptions and transactional customers we were spending a lot of time and effort monitoring the stock available through our website and app as well as the stock held back to serve our subscription customers. Initially we thought we could do better with a rule-based function managing stock on a schedule. Unfortunately this did not do the trick. Instead, we re-scoped the project and implemented a bespoke differential evolution algorithm and deployed it using AWS Lambda. It was an incredibly exciting experience seeing a project like this through to deployment within the first six months of starting my job.
What are you looking forwards to most in 2018 at Gousto?
Last year we went through initial deployment of a number of data products, including warehouse optimisation, stock optimisation, forecasting and our recipe recommendation engine. This year we’re taking each of these data products to the next level, introducing more advanced features, polishing performance and redeploying through a combination of Docker, Airflow and AWS Batch. I find it hugely rewarding to see our work evolve and I can’t wait to see what we learn along the way.
How would you describe the culture at Gousto?
As a company Gousto stands out because we’re an e-commerce business, we sell a subscription service, but have our own physical value-add processes in the warehouse, and change our menu on a weekly basis, essentially making us an FMCG business. At the end of the day we sell a simple product: fresh ingredients in a box with instructions on how to turn them into amazing home-cooked meals. Still, a lot of tech and research goes into making all this work, requiring teams with very different functions closely collaborate on a daily business. This shines through in a culture that rewards openness, inquisitiveness and ownership.
Why would you suggest that someone considers joining us in GoustoTech?
The way we consume food today is ripe for change. We’re taking a stand against food waste and for food to be enjoyed with friends, partners and family. Recipe kits are a catalyst for change in the food and groceries vertical, but in a way just a stepping stone to bigger things. Want to work on the engineering and data science to make this happen? Come join us.