We finish our #EdTechDataCareers blogs with this contribution from Michael Andrejczuk, Data Scientist at Featurespace. Read on to learn about what Michael has learnt both from his time at university and since starting at Featurespace.
Who am I?
I graduated in May 2022 from the University of Edinburgh with a BSc in Artificial Intelligence and Computer Science. I now work as a Data Scientist for Featurespace, a fintech scaleup based in Cambridge. In my role, I build machine learning models to help financial institutions to identify fraud and financial crime in real time.
What does my day-to-day look like?
No one day is the same for me! On a given day, I might be:
- Exploring a new dataset to see if it’s feasible for us to build a predictive model: this involves exploratory data analysis, usually in Python and SQL.
- Building a model: defining and testing features, tuning hyperparameters, training the model, and evaluating performance.
- Debugging a model that’s in production for a customer, looking through logs and plotting statistics to understand if it’s performing as we expect.
- Meeting with new customers to help them understand if our system is a good fit for their needs.
- Working with the commercial team to help them demonstrate the business impact of the models we’ve built.
The variety is one of the things I love about my role: it’s great to be able to work on a mixture of technical and commercial work. There’s always something new to learn and opportunities to pick up interesting projects.
The pros of working at a smaller company
When I was at uni, I thought I’d end up working at a large tech company: working at smaller companies just wasn’t on my radar, and all the internships I’d done were at companies that were household names. Since joining Featurespace, I’ve been surprised by how much I enjoy working somewhere smaller! I’ve been able to pick up lots of responsibility early on, and work on projects that are high-visibility across the entire company. These sorts of opportunities are trickier to get at larger, more bureacratic companies.
Finding something that excites you
I think it’s really important that you enjoy the work you do. When you’re looking for roles, think about what your values and motivations are, and how the companies you’re applying for align with that. The most important thing for me when looking for a role was ensuring my work would be making a positive impact on the world. The models we build help protect ordinary people from losing money to fraudsters, which is definitely a positive impact! It’s great to come to work in the morning knowing you’re making a difference.
What skills do I need to get into data science?
Technical skills are a must. Most employers will be looking for you to be comfortable with Python and key Python libraries used in data science (Pandas, NumPy, scikit-learn, etc). Having an understanding of databases and experience with SQL can also be helpful; the data you use at work will most often be held in databases, so you’ll want to know how to effectively query and process it. A lot of development work happens on remote servers like AWS, so experience with Linux and the command line will get you brownie points as well. Lastly, you’ll really benefit from having basic knowledge of software development principles: object-oriented programming, design patterns, unit testing, and version control. Demonstrating that you know not just how to write code, but how to write good code, will make your application stand out.
You’ll also want a solid knowledge of statistics and machine learning. You don’t need a PhD-level understanding, but you should know the basics: supervised vs unsupervised learning, the intuition behind common models, and statistical concepts like the bias-variance tradeoff and cross-validation. If you’re in Informatics, I would encourage you to take as many machine learning modules as you can. If you want to self-study those skills, then “An Introduction to Statistical Learning” is the best textbook to use: it covers pretty much all the knowledge you need for an entry-level role in data science, and has labs in Python to help you develop practical skills, too.
One aspect that’s often missed when discussing data science roles is the importance of understanding the commercial side of your work. Data science usually sits closer to ‘the business’ than a software engineering role would, so developing your commercial awareness and your ‘soft’ skills can really help make you a strong candidate! The careers service has plenty of resources to help you develop your commercial awareness, and company events (advertised through CareerHub) can be good opportunities to understand the market. Joining societies can be a great opportunity to learn public speaking and improve your leadership skills. Having technical skills is one thing, but knowing how to persuasively present your work is equally important, especially as you move up the career ladder.
Some finishing thoughts
Data science is a really exciting area to be in right now. Don’t constrain yourself to just large tech companies – lots of industries are exploring the applications of AI/machine learning right now, so you might find the role for you somewhere you don’t expect!
Featurespace regularly hire for entry-level positions in data science and software engineering, including internships, graduate schemes, and junior roles: https://www.featurespace.com/careers/
The Careers in Tech & Data Fair on 28th February offers the chance to meet around 40 organisations recruiting for a variety of tech & data opportunities. No matter your subject or year, this fair is for you – exhibiting organisations will have a huge range of opportunities for all. You can use the hashtag #EdTechDataCareers on MyCareerHub and socials to find employers with relevant vacancies and events. Drop by McEwan Hall between 12.30-4pm to explore! Go to MyCareerHub to find out more about the event including which organisations will be there.
