When musician Matthew Finch moved to New York and started managing a music venue, many of his responsibilities became data-related. Before he knew it, he had found a new passion in data analysis! Matthew explains why Flatiron School in New York City stood out from other data science bootcamps and how music and data science go hand-in-hand. Read Matthew’s advice about how to conduct a successful data science job search and how he landed his job as a Machine Learning Engineer at Bloomberg!
What inspired you to pivot from a career in the arts to data science and machine learning?
In college, I studied music composition and jazz bass performance. I got a record deal with a small label which I soon became a founding member of, and then moved to New York City to chase an artist's dream. Through a series of events, I became the manager for a music venue, and got more involved in the administrative side of the music industry. Without realizing it, I took on roles that incorporated data analytics for social media, scheduling, and time management. I came up with solutions to represent data visually, catalogued data, and tinkered around with SQL and Python. In my spare time, I was using creative software that ran audio through an algorithm. When I had a conversation with a software engineer friend of mine about binary search trees, he was confused because I didn't know how to code despite my knowledge of the subject. He suggested I check out a data analytics bootcamp.
There are so many data science bootcamp now — why did you choose Flatiron School in New York City?
I looked at a few bootcamps, but when I found Flatiron School, I appreciated their emphasis on helping students integrate their backgrounds with the tech industry. The Flatiron School graduates I spoke with found a job that incorporated their past career with their new tech skill set.
I also liked the energy at Flatiron School. They have a very New York attitude and understand the scrappy, ambitious individual who wants to take something and make it for themselves. I came from that world so I felt at home when I enrolled at Flatiron School.
How much coding experience did you need to know when applying to Flatiron School’s Data Science Bootcamp?
I was nervous and worried that I might not know enough math to apply to the data science bootcamp, but Flatiron School gave us a fantastic two-week warm-up that brought me through some data science and math basics. It also gave me a good introduction to how Flatiron School teaches.
Even in this prep work, Flatiron School not only teaches you, but they also give you resources. I went through all of their suggestions and found even more resources, creating a web of information. The prep work got me into a headspace where I felt confident about starting the bootcamp and beginning my data science career.
What was a typical day like at Flatiron School’s Data Science bootcamp in New York City?
I have fond memories because I loved my cohort. I used to get to the NYC campus 45 minutes early just to have breakfast and sit in the space. We would kick the day off with a coding challenge to crank up our brains. These challenges were hard, but I gave them my all. We would break off into sections that provided us with materials to work on individually or in groups. Then we had a lecture, another break, more work, another lecture, a second break, and work again! After the day was over my cohort liked to hang out afterwards. We might have a few drinks together, code, do some work, experience the community, and network.
Flatiron School encourages a mindset and a skill set that I use every day in my job and in my entire life. Flatiron School gets you from A to B, but gives you the tools to get from B to Z yourself. That skill is not as tangible as a technical skill like learning a website or algorithm, but it’s still a valuable thing that Flatiron School grads can rely on.
Flatiron School also gives you the time and space to ask questions and point you in the right direction to reach a conclusion. They don't hold your hand, but they also don't let it go either. It's an outstanding balance.
How many hours a week did you dedicate to the data science bootcamp?
Flatiron School asks you to dedicate 40 hours every week to the bootcamp, but realistically, everyone in my cohort who succeeded put in up to 60 hours a week. It wasn't because we had to, but because we wanted to.
What did you learn in the data science bootcamp curriculum?
The curriculum started with the fundamentals of Python, then SQL. After that we dove into micro-packages within Python, looking at Pandas and scikit-learn. From there, we covered statistical learning, math, and how to break down and implement concepts in Python. We learned how to do pipelines and retrieve data from online sources and compile their databases. We manipulated that data, built the statistical abstractions, and trained different statistical models. Flatiron School did a superb job of touching on the full spectrum of modeling, from simple ones to time-series analysis, natural language processing, and other deep neural nets.
Who did you learn with? Was your cohort diverse?
My cohort was made up of a lot of different backgrounds. We had people in the finance industry, humanities, political think tanks, and big box retail situations. There was a general average age range of mid-twenties to early thirties. No one in our cohort was beyond their late thirties, although other cohorts had older folks.
Flatiron School focuses on being a human, which my cohort encouraged to bond and become friends. I got more than an education from Flatiron School — I now have a network and a group of friends. We still connect through Slack and LinkedIn, and there is a general Flatiron School chat that we’re connected to. I also keep in touch with my cohort through social media, plus texts and phone calls.
What kinds of projects did you work on in the Data Science bootcamp and what did you build for your capstone project??
Coming from a music background, I focused on using music data in a lot of ways. One of my earlier projects was getting into Spotify's API and accessing their metadata to create a machine learning classification system that could identify and categorize different genres by looking at audio data. Other projects involved building a pipeline that connected from the back end of Yelp reviews, taking in the data, cleaning the text, and getting it prepared for a lot of modeling. There was a ton of data manipulation and extract transform load (ETLs) pipelines.
I was really excited to build my capstone project! I used MIDI and trained various models on Bach, Mozart, Debussy, and Chopin with an interactive front end. That way you could call the AI composer you wanted and, on the spot, they would play a 2-minute piano piece as well as give you the sheet music.
How did Flatiron School prepare you for the job hunt?
Flatiron School’s career prep was self study with break-out sessions. They prepared us to perform under pressure by pairing us together for mock interviews. They also taught us different ways of thinking about the information, to simplify and communicate it to others.
Flatiron School assigned me a career coach who kept me accountable. I was given a spreadsheet to keep track of where I applied, who I spoke to, etc. We were also encouraged to write and publish posts on Medium, and some of my blogs received positive attention, including one about my final project.
Flatiron School hosted a job fair, too. Since I was new to the data science career, I didn't know what it was like to interview for that kind of job. I had three interviews at the job fair, and two of them were great. The third one taught me who I wouldn't want to work for. It made me empowered to say no and find what I wanted. After graduating from Flatiron School, I was hired for contract work with a pharmaceutical company who needed a data analyst.
What roles did you feel qualified to apply for after graduating?
I never thought I would be a Machine Learning Engineer because I'm only good at data analytics and basic statistics. I soon realized a Machine Learning Engineer doesn't just do math and build models. There is a process to machine learning engineering that involves a lot of data analytics and data engineering.
I applied to every type of job and then braced for the wall of rejections. Every time I received a rejection, I would ask the hiring team for feedback and I learned a lot that way. It informed me of what I was good at and where I could improve. When I started realizing what jobs I could land, I analyzed the applications and the words they were using.
My advice to job seekers right now is – don't be afraid to apply and don't fear rejections. Learn from them. Each one can bring you closer to greater understanding.
Do you recommend that other data science bootcamp grads consider taking on contract work to help build up their resume after bootcamp graduation?
I think that it comes down to personal needs. I took contract work as a Data Analyst at a pharmaceutical company as an investment to make my LinkedIn and resume more robust and have more to talk about. When they brought me on, they told me I was the only one who knew Python in the company. That experience made me realize what I liked and didn't like and what I needed from any team that I worked with. I then did some contract work as a machine learning engineer with a cybersecurity firm.
That said, I'm just a single guy living in Brooklyn who has the time to do contract work. If someone has a family and needs to get a salary position, then they should put their energy into that goal instead. I had the luxury to take that detour and grab the experience.
How did you land your current job as a Machine Learning Engineer at Bloomberg?
I applied to Bloomberg three or four times for various positions. The recruiter contacted me and said that they noticed how I kept applying, and she wanted to keep in touch and talk. For around four months we occasionally chatted until one day she told me they had a job opportunity for me. She introduced me to a team, and after another 6 weeks, I was locked into a new job as a Machine Learning Engineer!
Did Flatiron School’s career services prepare you for a remote technical interview with Bloomberg?
My technical interview for Bloomberg included the exact same question that I had in one of my Flatiron School coding challenges! I did a three-part SQL series, and they asked me a natural language processing question to do text cleaning. Then they had me do data structure manipulation on the spot. It wasn't that difficult.
What are you working on now as a Machine Learning Engineer?
I work on a triage team that solves problems when internal technical issues are occurring. We are a small team that builds internal automation applications for the company to use internally. It's very back-end and internal, so there is no real glory, but it’s super nerdy and fun. I specifically do all the machine learning analytics and implementation of predictive modeling, which means I help to predict machine behavior internally.
Are you using the technical skills that you learned at Flatiron School?
The tools I learned at Flatiron School are the ones I'm using now on a daily basis. I've gone further into those programs and gotten deeper into AI and machine learning in my own time while using the Flatiron School foundation. Every tech job has a unique aspect to it. There are things that are idiomatic to Bloomberg that would be the same no matter where I worked. The land or domain knowledge was something I have to pick up, but what I learned at Flatiron School comes into play everyday like using Jupyter Notebooks, Pandas, and scikit-learn.
Are there any similarities between music and data science?
Coming from a music composition background, I often see the crossover between music and data science. Statistics and music work together. Music is a physical thing, so there is an objective reality to it, but there is a core mathematical aspect to how sound works, how we hear it, and the probability of each part leading to another. The data feels like I’m working on music: first the melody does one thing, then it interacts with this instrument, and becomes this other thing. Shifting over time, it's like listening to a beat that slowly changes, and by the end of the song, it's in a new place. There is a movement of information that becomes an expression.
I can actually deconstruct audio waves into data, and since I know that data really well, I implement it in different machine learning models to get different results. I am able to teach myself the next data science thing through that process by connecting the two. This process deals with the core mathematical reality of sound as a physical phenomenon and how probability and statistics work to relate events in time together.
Looking back on this career change, was Flatiron School worth it for you?
Completely! Flatiron School was entirely worth it. The structure, environment, and encouragement that came from the program kept me afloat. I see friends struggling in their life right now and I think that's where I would be if not for Flatiron School. I would never have had the accountability and five finished projects if I tried to teach myself data science. Flatiron School taught me how to talk about my trade and gave me the access to resources I wouldn't have known existed. If you are serious about making this career change, a bootcamp is a great idea, and Flatiron School is the best idea; it's an amazing choice.
NexGenT breaks down everything you need to know about networking certs!
Everything you need to know to prepare!
How FourthBrain helped Jorge land a job at AWS