Fun fact: Tyler used to be very fast! He competed in the UK School’s championships in the 400m relay and narrowly missed out on representing Ireland in the European championships.

What inspired you to join Vortexa? 

I really wanted to work on interesting, unsolved problems that allowed me to make the best use of my engineering and machine learning skills. At Vortexa, it’s difficult to be the kind of Data Scientist who doesn’t get their hands dirty with ML engineering and MLOps, and we have some of the most rich and varied data which allows for really creative modelling opportunities. This was exactly the kind of multi-skilled, cross-functional role that I was looking for after my previous career and academic experience and I have not been disappointed. 

What would you say is the most interesting aspect of your role? 

Understanding the intricacies of shipping! There’s so much complexity behind the scenes that us (previously) non-boat people really don’t appreciate. This role has demanded so much more domain understanding from me than any other role in my career – you genuinely need to sit down and learn how shipping and the energy industry works. And even then you’ll still find constant surprises!

What have been some significant milestone moments for you at Vortexa? 

I joined Vortexa as an experienced professional software engineer and with a little academic experience as a data scientist. In theory, I had all the skills I needed to do this job but there is a very large gap between those two disciplines that deals with putting models into production, professionally managing experimentation and results, presenting work to stakeholders and learning the intricacies of software design for machine learning specific tasks. The fact that I can now take a new machine learning project from conception, through requirements gathering and modelling and finally into production is a huge milestone in my career. 

What does day-to-day life look like for you as a Senior Data Scientist?  

A typical day for me (as a Senior Data Scientist who also likes writing production code) would look something like:

1. Decaf coffee (I’m getting old) and a nice walk around my neighbourhood (work from home flexibility is a blessing)
2. Daily standup to catch up with my immediate small team (pods as we call them)
3. Setting up other sessions on the back of stand-up to unblock myself or others where necessary
4. Reviewing Pull Requests for the pod I work in or other projects I contribute to in the wider team
5. Continuing with some exploratory data analysis or modelling that I was doing for a specific business problem (maybe I’m trying to improve predictions for what cargo is on board a vessel, or maybe I’m looking at our model that produces destination predictions)
6. Always lunch!
7. Either return to my modelling work, or use my results to implement some tasks that will run in one of our production pipelines. I like to have one data science task and one engineering task at once because the cycle time on the data science tasks can be quite long depending on training times etc. While a model is training and I await results I can be doing something else useful!
8. Usually a fairly high-level, blue sky meeting about improvements we could make to our processes, or interesting ideas we could try to use from recent papers, blogs or articles to improve our pipelines or models.
9. Reading! I try to read something useful every day. It’s all too easy to sit still in the worlds of machine learning and engineering and it’s a surefire way to fall behind! At the moment, I’m running a Statistical Learning workshop in the office that loosely follows the famous Introduction to Statistical Learning (sometimes known as “Baby” Elements of Statistical Learning by the same authors). It’s a great refresher for experienced people and an even better introduction for newcomers to the field. I’m reading a chapter of this book a week right now.
10. Several more decaf coffees in there somewhere.

What key qualities make a Data Scientist successful? 

Perseverance and hypotheses! When you follow online tutorials or toy examples for certain machine learning models it’s easy to get into a habit of always “solving” the problem, or finding a model that does the job very well. Real life (and indeed life on the high seas), is rarely like this, and it’s very, very easy to feel demotivated when results don’t come along quickly. You have to keep trying, keep brainstorming, and try the thing that you think “will never work” – sometimes it does! And hypotheses are important for similar reasons – when you’re learning they are often generated for you: “Investigate the relationship between engine size and car price. Is it linear?”. You need to exercise those hypothesis generating muscles to get good at this job and to make sure your investigations are always led by principled, methodical research and not random, spurious chance.

How would you describe the culture at Vortexa? 

The work culture is “curious”. There’s a curiosity to try new ideas, to try new approaches to working, to learn new things and to explore a question that hasn’t been asked before. If you are curious you will fit right in here!

What advice would you give to people hoping to join Vortexa? 

Ask all the questions. Every single one. It’s too hard to know everything and it’s too easy to forget what new joiners won’t know. Be prepared to ask lots of seemingly “simple” questions and to dig through a lot of data to find your own answers sometimes. It can be tough and overwhelming at the beginning but it does get easier.

Describe your working life at Vortexa in three words! 

Varied, Engaging, Boats!