Leaving Traditional Education to Learn Data Science


Sonali Dayal is a graduate of Galvanize’s data science program. Read on to hear how she followed her passion for real-world, practical skills in data science by leaving traditional education.

alumni sonali

What were you doing before the data science program? What prompted you to learn data science?

I attended Hack Reactor, a software engineering immersive program in spring 2014. I then enrolled in the UC Berkeley Master of Information and Data Science program. After three months at Berkeley, I withdrew to attend Galvanize, as I felt I wasn’t learning anything in the Berkeley program. The pace of the program was quite slow, and I didn’t have much of an opportunity to work with data in the first three classes. The program was focused more on theory than application. I was looking for the opposite, so I went to Galvanize to find that.

I’ve always been a fan of numbers and patterns. Data science seemed to be the new frontier, and I wanted to be a part of it.

How did you learn about the data science program?

I had been researching immersive programs for a career transition from biochemistry. I attended Hack Reactor, and I found that type of fast-paced environment to be ideal for learning. For focusing on data science, Galvanize was the obvious choice.

What’s the biggest misconception you had about data science?

Data scientist roles can vary widely. I had heard that before going through the program, but I didn’t fully understand to what extent until my job search. Many roles are very engineering-focused (e.g., data pipelines), while others are more focused on data analysis.

How would you describe a typical day at our data science bootcamp?

For the first 8 weeks of the program, we had a regular schedule of exercises and lectures. In the morning, we worked on brief “quizzes,” which were a review of previously learned concepts. After a morning lecture, we would work individually on reinforcing the topics covered. We then had an afternoon lecture, followed by pair programming for the rest of the day.

After 8 weeks of structured curriculum, everyone worked individually on capstone projects for 2 weeks. We had the opportunity to present our projects to hiring partners and prepare for the job search at the end of the program.

What’s been the coolest or most interesting project you’ve worked on so far?

I’ve been working on a lot of really interesting projects. At my current job, I’ve had the opportunity to contribute to Deeplearning4j, an open source project created by Skymind co-founder Adam Gibson. I’m currently building out a command line interface for Deeplearning4j, which has been really interesting.

I also have had the opportunity to work on building curriculum for Deeplearning4j training programs. I recently traveled to Seoul, South Korea, to assist with a training workshop at Samsung SDS.

What are you doing now? 

I’m a Data Science Engineer at Skymind. I get to work on a variety of projects, so my day-to-day varies quite a bit. Working at a young startup means doing more than one thing. A lot of my time is spent writing code, switching between Java, Scala, and Python. I’m contributing to Deeplearning4j, ND4J, and Canova, as well as using these tools for various client projects.

What do you love most about being a data scientist?

Data science has the ability to transform a wide range of industries. Particularly using deep learning, as I am at Skymind, the potential applications range from rapid, cost-effective drug development to creating self-driving cars. I love being able to contribute to that.

Any advice for aspiring data scientists?

Learn a little about everything. Know your stats. Write a lot of code (Python, Scala, Java). Don’t get complacent. Keep up with the newest techniques and tools out there.

Interested in becoming a data scientist? Learn more and apply for upcoming classes in Denver, Seattle, and San Francisco.