Taylor originally studied chemical engineering and was exposed to programming through the scientific research programming language Mathematica. After working in pharmaceuticals and therapeutics, Taylor realized that her background in engineering, statistics, and programming was ideal for data science. So she enrolled at Logit Data Science Academy in LA with the aim to get into a technical quantitative role in bioinformatics and AI. We spoke to Taylor about why she chose to study at Logit, what she has learned so far, and why she is excited about the data science industry.
What is your pre-Logit story? What is your educational and career background?
I received a master's degree from the University of Rochester in 2013 in chemical engineering. When I graduated I decided to apply my research-driven engineering education to biotechnology, so I joined a pharmaceutical company in Boston as a development engineer. I was a technical liaison between process development and manufacturing departments, which was fairly analytical. I was applying statistical and data analysis techniques to company business initiatives.
I then moved to the West Coast to work at the Emerald Cloud Lab. There I gained insight into a computer science aspect of therapeutics, and received training in computer science. There was a definite computer science element to understanding the technology, as the lab itself was powered by the Wolfram language. After I was exposed to this programming language, I decided to investigate how I could leverage my engineering degree and my statistics background with programming. I realized that data science is essentially rooted in computer science and statistics, so my background was actually very relevant.
How much programming knowledge did you have before this? Did you teach yourself about programming and/or data science? What types of resources did you use?
I had previous programming experience in mathematica through my engineering degree, but hadn’t actually implemented programming until I started at the Emerald Cloud lab. There I interfaced programming with advanced robotics.
I had also been teaching myself Ruby through MOOCs, including courses offered by Coursera, Datacamp, and Dataquest. I would definitely encourage all applicants to look at the online resources available, especially with Coderbyte and Codecademy. Also, there were a few textbooks, like Python Playground, and a variety of different resources you could use to understand what the scope of the course could be and what to expect.
Did you look at other data science bootcamps or just this one? What attracted you to Logit?
I was reviewing and comparing several bootcamp opportunities. A lot of them were online, and I actually utilized Course Report to scope out what opportunities were available. There are several other data science bootcamps emerging here in the market but their prerequisite requirements include a Ph.D. or extensive software engineering experience. So what you’re seeing in the market is around 40% of current data scientists have a Ph.D. according to recent reports, and the remainder have a bachelor's or masters. It’s not a requirement to get into the field, so Logit is actually reflecting what we are seeing in the market, and that is what predominantly attracted me.
I also liked the fact that Logit is a branch of an actual data consulting firm. And the founder is the host of Logit’s Hollywood Data Science meetup group, so he has networking connections, he knows what the field demands are, and what the responsibilities of a typical data scientist are.
Did you think about going back to college to study data science?
Right now, the market and demand for data scientists is surpassing what academia can support. The open source model, especially like you’re seeing in AI, is accelerating innovation of the tools that data scientists are using and expected to implement. There are a handful of masters and bachelor's programs in data science, but the scene is evolving so rapidly, that bootcamps are necessary to keep up with this technology. So bootcamps are doing very well because of the amount of skills you can learn in a condensed time period. The number of data scientists has doubled within the last four years, and there is a 50% higher rate of new data scientists than that of software engineers, so there is a huge demand.
What was the application and interview process like for Logit?
First, there was more of an informal conversation to discuss my expectations, what I was hoping to achieve, and my background. Students are required to have a degree in a quantitative field, programming experience and an understanding of basic probability and statistics. Then there was a more formal written application and a brief coding exam which I had a week to complete. There was no required language for the exam, you could code in any type of language that you knew, which was very appealing. I actually used Ruby, because I had been looking at web development, and I had been completing coding exams in Ruby. I have since been doing all my code in Python.
Was there a reason that you chose a data science bootcamp which taught Python?
Yes. There have been several studies which surveyed different tools that data scientists predominantly use in the market, and Python has come up as one of the most popular tools, so it’s a very relevant language.
What’s your class like in terms of size, diversity, gender, race, life and career backgrounds?
There are about 18 students and it is quite diverse. Students have a wide variety of STEM backgrounds including math majors, physics, CS majors, meteorology, and electrical engineering. So this particular course is attracting a diverse perspective. In terms of female:male ratio, there is me and one other female in the course.
What is the learning experience like at Logit so far — typical day and teaching style?
On a typical day there is a lecture and theory in the mornings, and labs in the afternoons. We have four professors who are rotating and teaching specific weeks of the curriculum. We are in the third week now, and the first two weeks we were taught by instructor Mike McKerns. He brought with him a vast amount of experience in the industry. He has authored 12 python packages, so we were able to ask him very specific questions about what he’s observed in his Python consulting.
So far we’ve been learning the fundamentals of Python, and Mike has shown us the back end of Python, how to understand the underlying C code of the functions, how to develop a pipeline of data analysis, how to approach solving large data problems, and how to streamline that workflow. Along the way, Mike has been using his industry experience to teach us these fundamentals and increase our understanding of the sort of real world problems we’ll be encountering. We’re eventually going to be covering machine learning and neural networking, which require these fundamentals.
Do you work on exercises or projects throughout the program?
We work on problem sets in labs in the afternoons, which reinforce some of the concepts we covered in lectures in the morning. Our course instructor teaches the lab, and we either work independently or in groups to problem solve and complete the problem sets. Towards the end of the program in weeks 11 and 12, we will have the opportunity to work on capstone projects, which will demonstrate our new skills.
What’s been the biggest challenge so far?
At this point, one of the more challenging parts is learning and understanding the new syntax at a very fast pace. The great thing about this course is we have such a diverse set of backgrounds, that people bring different strengths. My strengths are more in statistics and I’m able to leverage that experience. Everyone has different challenges, but we are all trying to learn at the same rate.
Could you give me an example of the sort of problem sets you’ve been working on?
A lot of the Big Data problems that people are encountering in the workplace are complex. We’re learning how to access public data through SQL, and in some of the examples we’ve been parsing data and giving summary statistics for particular subsets of this data. So we can find underlying statistical themes in these subsets, by performing analytical analysis with Python packages. We can also construct models by performing regression-based techniques on our data, which we will learn in the coming weeks.
A lot of these problems require a statistical approach, but we also need to be able to present them to clients. We’ll have the opportunity to hone our presentation skills during our individual mid-course projects.
What is the Logit campus like? Where is it and how big is it?
We are in the Broadway Hollywood building on the corner of Hollywood and Vine, it’s actually quite iconic. The Hollywood walk of fame is right outside our door, and on the rooftop there is a view of the Hollywood sign. We have lunches on the rooftop occasionally, as well as the meetup group event the last Thursday of every month. Classes are in a retrofitted loft.
Have you been able to give feedback about the program so far?
The director of the program and the founder have been excellent in collecting and receiving feedback. We’ve been filling out surveys, and they’ve been tailoring the program to the student responses. It’s really been very helpful.
How does the bootcamp prepare you for job hunting?
To ensure all the students are going to be finding jobs after this, Logit has partnered with a recruiting firm that specializes in data science. Towards the end of the course we will be refining our resumes, learning how to interview well and understanding how we can leverage our previous experience. Periodically throughout the program there are lectures from people in the industry. This week we have a data scientist from IBM coming to give a talk, which is going to be very helpful.
What was your goal in attending a data science bootcamp — get a job as a data scientist, start a business, get a promotion? What are your plans after you graduate?
I’d like to pivot into the biotech industry and do more of a technical quantitative role in bioinformatics. Especially in biotech and in pharma in general, you’ll start seeing that there is a cluster of opportunities to design experiments, find genetic markers for diseases, and other opportunities, especially in clinical trials, for determining how to select patients for those trials. So that is one potential implication of the skills I will be developing.
What advice do you have for people who are considering a data science bootcamp?
One of the motifs I’ve found is having the ability to adjust, adapt, and be creative. There are always multiple approaches that one can take to solve some of these problems. A lot of the data we’re going to look at is highly unstructured, so finding unique ways to implement the tools we are going to be learning will be extremely pivotal, and will definitely benefit future applicants to this program.
At this point the data science field is in its nascent stages. The name itself stems from a hybrid of a data analyst with a research scientist, and that came from Facebook’s first data group. So with a background in fundamental statistics, or computing, anyone can actually leverage these skills and apply them to this field. This is why data science is massively attracting people with quantitative backgrounds. It’s very encouraging and I would encourage anyone who has that type of experience to look into this field, especially people who are passionate about AI.