Nurturing the transition into data science: Louis Claxton, Faculty Fellow turned Data Scientist

During the first year of his PhD in Biogeochemistry, Louis discovered an interest in AI and machine learning during a university talk.

Curious about what machine learning was and what it could do, he spent the next three years independently studying the fundamental mathematics of machine learning and data science alongside his PhD studies, before joining Faculty as a Fellow in 2021.

“Studying a totally different field alongside my PhD made me realise that if you are dedicated enough, you can still do – and succeed in – data science, without a machine learning or maths background. You do need to put the groundwork in, but it is an excellent challenge. Don’t be conned by the ‘learn it quick or become an expert in 5 hours’ courses available online,” says Louis, who credits his online MIT OpenCourseWare introduction to linear algebra course for making all things data science ‘click’. 

Having a good foundation and understanding of core mathematical concepts was essential in beginning to understand machine learning algorithms. “Knowing the basics helped me recognise the roots of things – certain ML outputs and behaviours.” It was during this time  learning the basics that Louis began to question his direction, and he quickly knew that a move into data science was on the cards. 

Taking the first step

“Despite my PhD being in a different field, it offered me an opportunity to realise and explore a new interest in machine learning. It gave me the breathing room to recognise and question the area of work I wanted to be involved in,” says Louis.

It was during the second year of his PhD that Louis reached out to Faculty’s Head of Fellowship, Maria Diaz, equipped with his interest to learn more about data science. “I emailed Maria, asking about the skills I’d need to secure a place on the Fellowship. She kindly armed me with a list of resources ranging from heavy maths and statistics material to smaller project based activities. I completed this over the following two years, and soon enough, I applied to the Fellowship during the final year of my PhD.”

Louis discovered the Fellowship after searching for data science based roles outside of academia. Despite being offered a 4 year postdoc programme with his existing supervisor, he already knew that his decision to move into data science was final. 

“It was that two-year journey for me, having invested so much time during my evenings and weekends, that pointed me to the decision to leave academia. I had done all I could to make myself a great data science candidate, and it solidified my interest in the space. I enjoyed the potential it held, and I was drawn to working on projects with real world impact and short delivery timeframes. The postdoc position would offer me the chance to publish papers, but I wanted to produce work that demonstrated immediate value for clients. With the Fellowship, I knew this opportunity was out there.”

Making an impact

During his fellowship, Louis worked with Perse Technology, a startup data company looking to help people understand their carbon footprints in a more granular way. Focused on providing information to those looking to switch between energy contracts, Perse Technology helps its customers choose the most green and efficient tariffs available.

Louis’ project was, as he terms, “a classic customer segmentation problem”. He used a form of unsupervised machine learning to help the company understand who their customers were and why, and he spent the duration of his project clustering customers to identify what categorised each customer ‘type’. Louis’ work helped Perse Technology best determine how to target customers via personalised marketing, and how to treat them to suit each specific customer need.

“My Fellowship project taught me that nothing really prepares you for dealing with real life clients,” says Louis, who noted the jump from working mostly independently in academia. “You can do all the prep you want, but nothing will change the fact that you might encounter a change of project scope on day one, or need to deal with missing data. You have to tackle messy data, client management, and find new ways to approach a problem.” 

For Louis, this was an experience that he felt he couldn’t get elsewhere. “There’s a team, project curveballs, plus the day to day client management. It’s not just sitting in front of a computer and coding.”

A data science career

After graduating from the Fellowship, Louis now works as a Data Scientist in the Government team at Faculty. “The Fellowship really did prepare me for what data science looks like in the wild,” remarks Louis, who now works on a range of projects with ML applications in the fields of anti-terrorism and refugee relocation. “I know I’m not going to get perfect, clean data sets from the web and just write machine learning code. In reality my role covers a lot of data engineering and building out pipelines before getting onto machine learning.”

Louis says that the Fellowship gave him not only machine learning skills, but the data engineering skills to hit the ground running in his first role. His best advice for prospective Fellowship applicants? Keep learning and exploring your interests, and reaching out to previous Fellows to get their insights. “Since joining I’ve had applicants reach out to me, asking about the Fellowship. People will be willing to give you five minutes, or point you in the right direction. For me, the fellowship has been brilliant, and I’d certainly recommend it.”

To learn more about the Faculty Fellowship, click here.