When Tiffany Li heard about the power of machine learning to automate tedious processes, she turned to Metis’ data science bootcamp in Chicago to transition careers from consulting to data science. Today, Tiffany is a Data Scientist at Grubhub in Chicago, solving a variety of business problems alongside other Metis alumni. Tiffany tells us about the Metis application process, the difference between learning at Metis versus college, and the demand for data scientists in Chicago (spoiler alert: Metis grads are working at Amazon, Facebook, Airbnb, BuzzFeed, and Spotify)!
Can you tell me about your career and education background, and how that led to you to study data science at Metis?
I graduated from Northwestern University in 2015 with a double major in Economics and Math, but I didn’t take many coding classes. After graduating, I worked at Deloitte in Chicago doing consulting work with clients from different industries – I did qualitative research and some financial analysis work with Excel. That was all very interesting and I got good exposure to the business side of different industries, but I wanted to advance my technical skills.
I started researching data science and I thought it was fascinating that so many companies were using machine learning to automate processes and make better decisions. I wanted to learn to code, and to learn about machine learning so that I could get to the next level and use more advanced tools like Python in my job. That's how I discovered Metis.
Since you already had a background in math and analysis, did you consider teaching yourself?
Yeah, I actually started by taking online courses. I checked out Coursera and completed a Nanodegree on Udacity. I started learning the foundations of Python on my own, but I needed more instruction and motivation. Learning on my own was hard because every time I ran into an issue, I had to spend a lot of time researching online, which slowed me down and made me less motivated. I wanted an option that was more intense and would provide more support so I didn't have to waste so much time Googling.
Why was Metis the right bootcamp for you over college or other bootcamps?
After checking out those online courses, I did a seven-day bootcamp in Chicago called Data Science Dojo. That was helpful to learn some basic machine learning ideas, but wasn't long enough for me to get my hands dirty. I wanted a longer, more intense program.
I thought about getting a master’s degree, but I felt like the whole process of taking the GMAT, applying, and going to classes would take too long and be too expensive. The three-month bootcamp option was the perfect combination – with both the support I needed, and a short enough length that I didn't have to spend years studying to get into the data science field.
I came across Metis in a blog post about different types of data science bootcamps, and I noticed a lot of them required PhDs. Out of the ones that I was eligible for (without a PhD), Metis was one of the best, based on reviews and their alumni placement. I was originally planning on going to Metis in San Francisco or in New York, but when I heard that Metis was going to open a Chicago campus, I was really excited that I could do Metis without having to move. I applied and I enrolled in the first cohort in Chicago.
What was the application and interview process like for Metis?
For Metis, you need to have a sound background in Python, math, and a basic understanding of data science, which is great because it means the instructors don't have to spend much time teaching the foundations.
The application process is actually three stages:
It was actually a pretty rigorous process – which is a good thing because it ensures that the student quality is high and that everyone has different skills they can bring to the class.
What was your cohort like? What kinds of backgrounds did the other students have?
My class was made up of people from backgrounds in finance, consulting (like me), engineering, and web development. There was a wide set of skills and everyone could help each other in different ways. For example, if I had a question about a front-end app, I would go to a friend who used to be a web developer because she has that expertise. My background in math, so I could help other students with probability problems.
We were lucky that our cohort was very diverse in terms of gender. We had nine students in total and four were female students. We also benefited from the $3,000 scholarship that Metis gives to women and other minority groups.
What was the learning experience like at Metis? Can give me an example of a typical day?
The three-month bootcamp was pretty intense. We had class Monday through Friday from 9am to 6pm. We started each day with pair programming. Then we would have a morning lecture covering things like supervised machine learning, unsupervised machine learning, probability, statistics, logistic regression, and neural networks. The mornings were full of new information that you had to absorb quickly.
In the afternoons, a student might present a data science-related topic that they researched, or we might get to work on our individual projects. We worked on five projects during the three months – one group project and four individual projects. We got to choose topics that we were interested in, gather data, do analysis, and present to each other. The projects were really helpful because we could immediately apply the knowledge we learned and that helped us to reinforce our learning.
How did the learning experience at Metis compare to when you were studying at college?
It’s a different learning experience. In college, the classes often focus more on theory. For example, you might learn about a great mathematician who developed a theory and how to prove it, but not about how the theory applies to real life.
Whereas Metis is very practical. They do a great job of customizing the curriculum so that whatever you learned in the bootcamp can be immediately applied to your job. Metis would change their curriculum based on students’ feedback, and invite speakers who work in data science to talk to us about what kind of tools they use. You can apply whatever you learn at Metis, not only to a job but also to job interviews.
How did Metis prepare you for those job interviews?
Each cohort has a full-time career advisor. Our career advisor in Chicago – Ashley – was really great. She did workshops about how to polish our LinkedIn profile, how to network, and how to negotiate salary. She also brought speakers from the industry, which could potentially lead to an interview. After you graduate, the career advisors help you navigate the recruitment process, help you with mock interviews, and get you connected to employers. That's really important.
At the end of the bootcamp, Metis holds a career day where students present their final projects. They invite employers who have data science job openings, to see the presentations and talk to the students afterward. A lot of times that would result in interviews immediately after the bootcamp, which was really great.
The other huge advantage of Metis is the alumni network. A lot of alumni are working at great places like Facebook, Airbnb, BuzzFeed, and Spotify. In Chicago, a lot of Metis graduates are working at GrubHub, which is how I found my current job. A Metis grad (who studied in NYC) saw my presentation and he knew that his team at Grubhub had an opening. He referred me and got me into the pipeline quickly.
Did you see a big demand for data scientists in Chicago while you were job hunting?
There is a lot of demand in Chicago. It's definitely not as huge as New York and San Francisco, but there's also a smaller supply of candidates. When Metis first opened in Chicago, there were a lot of companies very eager to hire data scientists. There are a lot of opportunities, but it can be tricky because many companies want to hire someone who has a little more experience. I was lucky to find this opening and get the job at GrubHub.
From our cohort, my classmates are now working at huge companies, but also at startups and medium-size companies. People got jobs at Amazon, at trading firms doing quantitative research, or at computer vision data science companies. Others focus more on the data visualization side of data science or are working as Cloud engineers – everyone ended up working in really good roles that they're excited about.
Can you tell us more about your role at GrubHub? What’s it like to be a data scientist??
Right now I’m on a team with two other data scientists. GrubHub doesn't have a centralized data science team so the data scientists are placed across different departments. For example, we have data scientists in customer retention, operations, and in research groups where they are focused on longer-term research projects. The good thing about that structure is that if any team has questions, they always have a data scientist on hand whom they can go to.
I work on a variety of different projects, but I spend a lot of time with our product team. If the product managers want to test out a new feature on our website or in an app, I help them figure out which metric to measure, then we monitor the results over time with A/B testing. Or a product manager might want to know why their new feature failed, so I can look into the data to help them figure out where people are dropping off, and what to do next.
The other part of my role is working on research projects. For example, figuring out the impact of weather on our business (GrubHub gets more orders when it's raining). I also look at the impact of changes to delivery fees or how we design our review system. There are tons of interesting projects and a lot of variety in my job.
Are you using the same technologies and languages that you learned at Metis?
At Metis, you learn Python and SQL – I use both in my job. Then there are other parts of the job that are more specific to the role. For example, we work with ETL (extract, transform, load), which is designed for automating the process. At Metis, you learn a little bit about Git, but I use Git a lot at GrubHub. In general, Metis covered most information, but you'll never be able to learn all the specific software that each company uses. As long as you have a good foundation of Python and SQL, you will be able to learn new technologies really quickly.
How did Grubhub support and train you when you first joined?
The training process is very hands on. I spent a lot of time working with other data scientists to understand our legacy code. We also had meetings with data scientists from other departments to know what they're working on and to see if there's any overlap in our work that we can collaborate on. GrubHub also lets us attend data science conferences and supports us in meeting people outside of the company to learn what's new in the industry.
Since you graduated from Metis and joined GrubHub, how do you feel you've grown as a data scientist?
The biggest thing I’ve gained is understanding the business. In addition to having technical skills, data scientists need to understand a business to know how to apply those skills. For example, the data might be saying one thing, but if you have some business context you’ll realize the data could be indicating a larger problem.
I also have a much better understanding of how other technology companies are utilizing data science and the different parts of data. As well as machine learning, I’m also seeing the importance of data engineering, data visualization, and how you present your analysis and convince your stakeholders. Business knowledge is super relevant for data scientists.
How useful has your background in math and working at Deloitte been in your new career as a data scientist?
There are a lot of small things I took from my past career such as how to draft an email, and effectively communicate ideas, and how to manage your time and organize all your tasks. In math, that's really important because we need to be very rigorous in our analysis. We need to be careful about the pitfalls that we might fall into if we only look at the surface. We always need to validate our ideas and make sure that we can back up our analysis.
How have you been able to stay involved with Metis?
I try to go to alumni events and reunions and Metis meetups in Chicago. I also like to attend other cohorts’ career days, not only as a grad, but sometimes as an employer. It's good to go back and check out the newest tech out there – I always learn something new from the new Metis grads.
Since I started at GrubHub, we have hired a couple more Metis grads – one is working on the research team in our Chicago office, and the other is working on my team, but from New York. We are always bringing more fresh Metis grads into the company.
For people in Chicago thinking about going to data science bootcamp, what’s your advice?
First, it's really important to research the bootcamp. After doing a lot of research, Metis came across as the best one because they care a lot about the students and they put in a lot of effort and support. I recommend going to a website like Course Report to read students reviews.
Go to bootcamp info sessions to get an overview of what the bootcamp teaches. Sometimes bootcamps will host events like "One Day at Bootcamp" where you can get a feel for whether you’re interested in data science, and if you can handle the intensity of the bootcamp. So go to a lot of meetups, talk to instructors, alumni, and current students. That's important before committing yourself to a three-month bootcamp – it's not for everyone.
Is there anything else that you want to add about Metis?
The Metis community is really great. The staff and instructors are always supportive. You have two full-time instructors, one full-time career adviser, and one full-time program manager. So you get a lot of undivided attention from the staff.
Another great thing about Metis is they're very open to feedback. For example, they invite alumni back to hear about what we're working on and they ask us if there are any technologies Metis should add to the curriculum. Right after we make suggestions, they take action and change the curriculum accordingly. Seeing feedback get implemented is really great.