Career Series: Life as a Data Scientist
What does a day in the life of a data scientist look like? What skills and mindset do you need to break into the data industry?
As part of our Career Series, we asked Anabel — Data Scientist at Urban Logic — and Georges — Applied Machine Learning Scientist at Microsoft — about their journey getting into data science, their day-to-day, and their tips and tricks for those who want to follow a similar career path.
Can you introduce yourself and tell us about your background?
Anabel:
I’m a Data Scientist at Urban Logic. We help local governments analyze and visualize all kinds of data from traffic lights to real estate education.
I have a Ph.D. in Physics. I used to work at the CERN on particle physics experiments, and from there, I moved into the data industry, working for startups focused on social impact.
Georges:
I work at Microsoft as an Applied Machine Learning Scientist for the Azure text analytics team. We build natural language processing algorithms, meaning we teach the computers to understand and generate text.
How did you get into data science?
Anabel:
When I finished my Ph.D., I wondered how to use my skills to solve problems, and data science presented itself as a natural choice for me.
It helped that I already learned part of the required skills during my Ph.D. For instance, I developed an algorithm to sort elements in a detector. It was similar to what you could do as a data scientist.
Georges:
I studied Mathematics and worked in business consulting in Mexico before coming to Montreal for my Bachelor’s final project. My supervisor was doing his Ph.D. in Machine Learning, so naturally, I came to work on that as well.
After my studies, I landed a job as an AI Researcher for a Montreal-based EdTech startup. I worked on developing deep learning algorithms. It was the perfect opportunity to get my hands on data science projects. Less than ten months later, I was contacted by Microsoft for an interview, and before I knew it, I was flying to Seattle to start a new job as an Applied Machine Learning Scientist.
What are some data science applications you work on?
Georges:
I work on Sentiment Analysis to teach the computer how to recognize a positive, negative, or neutral tone. I also work on Summarization algorithms to highlight the most important parts of a newspaper article or a book. The models I worked on are then integrated into the cloud for our clients to use.
Anabel:
Right now, I’m working on an Origin Destination Predictor. We use data from location and trackers to understand how people move in a city. For instance, if the city needs to block a road for construction, we want to predict how the traffic will dispatch itself.
What tools or technology do you use on a daily basis?
Georges:
I was looking at Le Wagon’s data science syllabus and, actually, I use a lot of the tools that you teach in your programs.
For instance, I use a lot the machine learning libraries from Python like TensorFlow or PyTorch.
For code editing, I use Visual Studio Code, and for model training and dataset storage, I use the Azure cloud.
Anabel:
I use Azure as well and Sublime is my favorite code editor.
I use Python but I also learned JS and VueJS at my previous job. I think that the ability to understand the front-end and back-end of web applications is very helpful.
What are some of the challenges you encounter at work?
Anabel:
It’s funny because we were discussing this with my team. We’re currently reviewing how we manage our workflows and we realize that the agile task-based approach might not be the right fit for us.
When you work in data science you work on issues that you don’t really know if it’s going to work. We operate a lot by trial and error, and we can’t always clearly define tasks and goals and check the boxes.
I think the biggest challenge is to figure out the best way to streamline workflows.
Georges:
Working with code can be frustrating sometimes. You know you have a problem and you don’t know where it is. Of course, you can get help from your team, but you’re the one who owns the project and has to make it work. So it’s important to have patience.
But also remember that Stackoverflow is here to help you 😉
What soft skills do you need to be a successful Data Scientist?
Georges:
At Microsoft, we value the concept of growth mindset. So instead of having a fixed mindset and trying to look smart, you should always be asking questions and be curious.
When you’re not sure about something, ask questions and look for feedback! That’s how you’ll grow as a data scientist.
It’s also essential that you show your work to other team members, even outside of your team. That’s how you’ll build social relationships at work and learn from your peers.
Anabel:
Persevering is the key! Trying different things and exploring new concepts is part of the job of a data scientist.
You should not be afraid of not knowing. As Georges said, it’s important to ask questions — not only to people in your field but people outside of your field. As a data scientist, you need to consult with experts in the field you work on.
I’ll also add that being able to share your knowledge and your discoveries with others in an understandable way is a valuable skill. For instance, with your business team or customer support team. If people can understand what you’re doing, they’ll be more prone to ask the right questions, which is how a company makes progress.
What would be your advice for junior profiles who want to break into the industry?
Georges:
Experience is always required for any job, even an entry job. So, you can always start with an internship or work on your own project. If you can afford to be paid as an intern, it’s a good strategy because you’ll build up some experience and then you can either stay in the company or move on to something else. You can look at the co-op opportunities for instance.
Also, If you have connections in the industry you want to work in, it’s important you reach out because they will make you advance faster in the process.
And then apply to as many interviews as you can to be able to practice live coding! Failing at interviews is the best way to learn.
Anabel:
I agree. Internships are very well valued. I can see it in the company I’m working in right now. We see dozens of resumes every week and being able to state co-op introductory experience is super valuable. So I would definitely encourage that.
If you see that it’s difficult to find a job, start small, and after a few months, you will upgrade!
Thanks to Anabel and Georges for sharing their day-to-day as data scientists!
If you’re interested in learning Data Science, we invite you to take a look at our bootcamps:
