Physics Undergrad to Data Scientist

Callum – Ultimaker Data Scientist (LinkedIn)


Tell us about university. What did you study and what did you do after graduating?

I did four years of physics and astronomy at uni. When I finally graduated, I found a summer job doing manual labouring just to earn a bit of money while I figured out what I was aiming for in the slightly longer-term. I spent a couple of months doing that before I secured a job as an entry level litigation support analyst (i.e. a data analyst) for a legal firm. The idea of working in tech held some appeal and I thought this would be a decent way to get my foot in the door. Although I’d have never guessed my first job out of uni would be for a legal services company!

What kind of work did you do as a data analyst?

Lots of product-focused work. Some of the technology products that the company sold would do things like classify legal documents using machine learning. Whilst it was never my job to look deep under-the-hood of the algorithms behind this software, I did want to learn more about the work of a data scientist so I made the effort to spend some time learning to tweak parameters and optimise the models. But the majority of my other work involved using a SQL-based platform to work with large datasets, creating reporting dashboards for management, and communicating with clients.

How did you move from data analyst to data scientist?

For personal reasons, I ended up moving to the Netherlands and finding another job there as a data analyst in an English-speaking company. The company is all about 3D printing which is generally a pretty innovative sector, so I was very happy to get the role. Again, whilst doing my work as an analyst, I’d take every opportunity I could get closer to machine learning and data science in general. I also did several machine learning training courses online, both on company time and in my own. After some internal changes, we needed a data scientist and I proposed myself for the position, given my previous experience and familiarity with the setup.

The path I had to create from litigation support analyst to data analyst to data scientist taught me that there’s not really such a thing as a junior data scientist, for which there’s very few listings on job search platforms. Since it is such a great role to have, data scientist jobs are highly competitive and so companies can afford to pick out candidates that already have direct experience or some experience in a similar role, such as a data analyst who’s familiar with Python and machine learning. My physics degree was definitely important in getting me the role but it certainly wouldn’t have been enough alone.

What training did you do?

I did some of the DataCamp courses. The ‘Data Scientist with Python’ course I especially liked. It’s not free, but I thought it was good value for money. The important point though, is that these courses, or whichever ones you do, are only for learning. They miss the practical application. For example, you’ll never get the experience of having to clean a dataset before you can actually do machine learning on it if you only follow these courses. You’ve got to go and apply these tools and techniques to some kind of mini-project of your own. I was saying earlier how data science roles are highly competitive; showing that you’ve spent time actually applying whatever tools you’ve learnt will be really important in setting you apart.

So what tools and technologies are you using most often as a data scientist?

My company has embraced Google Cloud. So I’ve become familiar with loads of the products on there; Data Studio, Cloud SQL, Cloud Storage, etc… We use Power BI for reporting occasionally but to be honest, more people are embracing Data Studio every day. For specifically machine learning, Google Cloud offers loads of pre-built and customisable models which are worth getting to know. Sometimes I use Python for when I need something very specific. And I’m thinking about learning R, but not sure how useful that will be right now. The whole field of machine learning and data science is changing so fast that to me, it is more important to learn how to apply the algorithms in a business than to know all the tiny details and complex maths of exactly how they work, especially as AI services are becoming more and more automated.

Do you enjoy your job now?

Yeah absolutely! AI is such a fast-growing field and I’m definitely happy to be a part of it. Admittedly, this can sometimes be difficult because there’s always so much new stuff to know but overall, this is the place I want to be. I also think working in the Netherlands suits me really well. I tried working in London but the constantly high pressure, non-stop client-facing work was not for me. Here, they really respect the work-life balance and it feels like a great fit culturally for me!

Any tips for wannabe data scientists?

Yep.

  1. Like I said earlier, having a numerical background and a machine learning course under your belt is not really enough to get you a data science role. You absolutely need the application experience as well, whether that is from projects you do in your own time or from another similar job role.
  2. Be ready to be a data evangelist. If you’re not sure what that means, look it up. But in short, it is someone who is prepared to explain and sell the idea of data to usually non-technical colleagues. You won’t always be surrounded by people that understand the value that data science can bring!
  3. Don’t be afraid to specialise. This might be a surprising thing to tell someone new to data science, but as the field has grown so quickly over the last decade, what one data scientist does compared to another can be massively different nowadays. So if natural language processing interests you, go ahead and get really good at that. If creating machine learning pipelines on the cloud is for you, then great, focus on that. Some breadth of knowledge will always be advantageous but the demand for generic data scientists may diminish as the market grows and saturates, with organisations able to afford to seek specific skillsets.

Views expressed in guest posts are solely those of the writer and do not represent the thoughts, opinions or views of any other mentioned third parties, including employers or colleagues.

Share:

Leave a Comment

Your email address will not be published.