From chemistry labs to leading a Data Science team — how curiosity and a love for learning got me there.
My path into data science wasn't linear. I didn't have a computer science degree, a portfolio of side projects, or a network of people already in the field. What I had was a physics background, a genuine curiosity, and a willingness to figure it out as I went.
This is the story of the six steps that actually moved the needle for me — and that I believe can work for anyone.
The temptation when starting out is to memorise syntax — to learn Python or SQL as if they were foreign languages with vocabulary lists to study. That's the wrong approach.
What matters is data fluency: the ability to query data, write clean code, and communicate insights. Focus on Python, SQL, and data visualisation not as tools to master but as languages to think in. The goal is being able to move from a question to an answer, not to recite functions from memory.
I became fascinated with recommendation systems after seeing a BBC job posting that mentioned them. That was it — I fell down the rabbit hole of collaborative filtering, content-based recommendations, and the ethics of algorithmic decision-making.
I completed Frank Kane's course on building recommender systems. I read papers. I experimented. That obsession gave me something to talk about in interviews that felt genuine rather than rehearsed — because it was.
You don't need to know everything. You need to know something deeply, and care about it visibly.
Employers in data science aren't primarily looking for people who know the right answer. They're looking for people who can figure it out. The ability to connect data work to human outcomes — to ask "so what does this mean for the user?" — is what separates good analysts from great ones.
In interviews, demonstrate initiative and learning ability. Show that you've taken imperfect information and done something useful with it. That mindset is more valuable than any specific technical skill.
I gained production experience at BBC working on recommendation engines. I learned deployment, Docker, CI/CD, and cloud tools at Trustpilot. I discovered a passion for explainable AI and SHAP frameworks somewhere in between.
Nobody enters their first data science role knowing everything. The best companies know this and hire for trajectory, not completeness. Your first job is where you learn to be a data scientist. Your interviews just need to show you're worth investing in.
Technical skills open doors. Soft skills build careers.
Curiosity, communication, empathy, and the ability to collaborate across teams — these are what distinguish people who rise through data science from those who plateau. Data is only valuable when it informs decisions, and decisions involve people.
The best data scientists I know are as comfortable in a room full of engineers as they are presenting to a board. Work on both sides of yourself.
Coming from physics, chemistry, or another quantitative field is a genuine advantage — but only if you package it correctly. Recruiters won't automatically see the connection between your dissertation on thermodynamics and a data pipeline job.
You have to make the translation explicit. In your CV, in interviews, in how you introduce yourself. Frame your existing skills in data language. Highlight the problem-solving mindset, the comfort with uncertainty, the ability to design experiments. That background is a differentiator, not a liability.
Breaking into data science doesn't require a tech degree. It requires sustained curiosity, a practical orientation, and the willingness to develop both technical and interpersonal skills — while solving real problems.
The field needs more people who think differently. Your unconventional path might be exactly what a team is missing.