Classical guitarist Matt Linder spent a decade as a performing musician and teacher, but when the pandemic canceled his gigs, he took the opportunity to follow his interest into tech and machine learning. After spending a few months of dedicated self-teaching, Matt enrolled in FourthBrain’s Machine Learning Program to level up his knowledge. Matt shares how FourthBrain’s part-time program empowered him to learn machine learning and launch a new career as a Machine Learning Engineer at Headspace!
What inspired you to make a career pivot into machine learning?
I finished my master’s in classical guitar performance at the San Francisco Conservatory of Music. For the past decade, my time was split between touring and teaching, and I started to hit major burnout working as a freelancer. At age 28, making a living in San Francisco – the most expensive city in the world – as an artist felt pretty cool. At age 32, it wasn’t quite as shiny and new. I realized I needed something else.
I have always been a huge fan of science fiction and concepts about artificial intelligence. I’ve been fascinated by the underlying processes and concepts at the core of machine learning and AI, even if at the time I didn’t know the term “machine learning.” Working with my bandmates, as our lives started to get more technologically focused, I wanted to see if we could use machine learning. I kept having musical ideas that were inspired by machine learning and AI, but I didn’t know how to make them work. One of my bandmates graduated from Holberton School and became a software engineer, so I would bounce my ideas off him.
Why did you choose FourthBrain?
I had always done what I loved, so at first it was difficult for me to consider working outside of music. I did a lot of self-study and started considering bootcamps, but I wasn’t sure how I could do a full-time, 12 week bootcamp and work at the same time.
In a machine learning newsletter I subscribe to, Andrew Ng mentioned FourthBrain, a program that his fund was backing. The fact that Andrew Ng backed FourthBrain meant a lot to me, so I checked out their program. FourthBrain’s program is part-time — plus, the tuition was affordable and the curriculum looked amazing, so I applied!
When looking for bootcamps, was your goal to become a machine learning engineer?
By the time I decided to do FourthBrain, I knew enough to know that job titles in data science would vary depending on the company. I knew I couldn’t get too attached to any specific job title, so instead I kept my eyes on the type of work I wanted to be doing.
Did FourthBrain offer any financing or payment plans to make the tuition more affordable for you?
When I reached out to FourthBrain, I mentioned that I was a musician interested in financial aid. The FourthBrain team was immediately responsive and willing to be flexible. I received a scholarship and was able to pay the remaining tuition in installments.
What was the application process like for FourthBrain? Was it challenging for you since you had been self-teaching?
FourthBrain includes an assessment in the application process, and I found it reasonably challenging at the time since I came from a background as a professional musician who had been doing self-study for a couple of years. For someone with a software background, the assessment would be a light version of a LeetCode-style interview combined with a light test on probabilities, statistics, calculus, and linear algebra. For a lot of tech people, the assessment wouldn’t be a burden. For me as a musician it was a nice challenge. That said, if I was able to do it, anybody who’s able to resource themselves properly should be able to do it.
Did you have to complete any pre-work?
FourthBrain had us complete pre-work that was “Week Zero” for the 16-week program. In Week Zero, there was a review of the concepts from the assessment to make sure you’re up to speed for week one of the bootcamp.
Was it possible to work part-time and also complete the FourthBrain bootcamp?
It’s totally possible to do the bootcamp and work part-time. My bootcamp experience happened during the pandemic, so my touring/performing work was more or less gone. Luckily, I was able to teach up to 20 hours a week. I think working part-time gave me an advantage in the program because of how much time I could put into the program compared to my peers who were working full-time. The bootcamp is a rigorous and challenging program, and it’s the kind of work that you put as much time into as you have available. It’s clearly doable to work full-time while completing the bootcamp, but sometimes my peers that were working full-time and/or had kids seemed pretty taxed.
What was a typical week like in the FourthBrain bootcamp?
We had class for six hours on Saturdays, and then during the week, we had about 15 to 20 hours of coursework to complete. Class-time on Saturdays was intense; there was a combination of lectures, practical projects, group projects, breakout rooms, and working on problems with various people before finding our final capstone project teams. There were extra credit projects after each class session. I put a lot of time into a couple of these because I wanted to be at the top of the leaderboard. After all of the lectures and assignments, you can always go deeper in learning these concepts — I was never bored or at a loss for more things to study.
What did the Machine Learning Program curriculum cover?
During the first four weeks, we covered machine learning basics, such as supervised learning, regression and classification, unsupervised learning, and semi-supervised learning. After that, we covered four weeks of deep learning. The deep learning unit had a heavy focus on Computer Vision because our instructor is a leader in the Computer Vision world. Once we hit week eight, we had started to form our capstone project groups. There was a midterm, and then we started on our projects.
The midterm was hard for me. I put a lot of time into it and I got a 100% on it, which felt really great. That was key for me since I felt imposter syndrome coming from a different background outside of tech.
While the capstone projects were going on, we were delving into advanced concepts in deep learning. We also covered machine learning deployment, MLOps, and other things that synergized well with our capstone projects. In the final week, we had a presentation for our capstone projects.
What did you build for your capstone project?
First, each of us pitched two ideas. If we could get other people to sign on, then it would become the capstone project. I came up with project ideas that were based on my career as a musician, audio engineer, and content producer. I was lucky that a lot of people wanted to join my group to work on these projects with me, and that made me feel good because imposter syndrome is very real. Being the only musician in a group of software people, I didn’t feel like I was the most-learned a lot of the time, but I was glad people liked my ideas.
For my capstone project, we built a speech enhancement project. The concept was to develop a deep learning algorithm that could automatically enhance recorded spoken audio. Imagine you want to record a podcast with your friend or start your own YouTube channel, but you don’t know how to make your audio sound like a pro does. We created a deep learning tool to automatically take any sort of spoken audio recorded on consumer devices and enhance it so it sounds like it was recorded in a professional studio.
What was your cohort like?
The cool thing about FourthBrain’s online bootcamp was that we had people from all over the country in our cohort. My cohort was about 25 people, and it was a cool, diverse group. Many people in my cohort worked in software in some capacity, a few people were already data scientists, and there were plenty of software engineers. There were also a few other people like me that weren’t as experienced in the tech field.
How did FourthBrain help you prepare for the job hunt in Machine Learning/AI?
Career services at FourthBrain is led by Mariam, who is awesome. She gave us ample help in terms of updating our resumes and interview prep, and I still correspond with her because she's a great resource. There were events with recruiters and Mariam organized on-on-one informational interviews with people in the industry for each cohort member. There were also industry expert guest speakers who we could ask questions and get advice.
After graduating from FourthBrain’s Machine Learning Program, which roles did you feel qualified to apply for?
You can’t always rely on job titles, so you have to pay more attention to the job description. I’m working as a machine learning engineer right now, but I also interviewed for positions with titles like data scientist, machine learning scientist, machine learning research scientist, and natural language processing engineer.
What is the difference between a data scientist and a machine learning engineer?
Typically in the industry, data scientists are charged with developing statistical and decision-making understanding of company problems and then developing models to understand or solve those problems. Machine learning engineers are often responsible for designing, building, and maintaining the infrastructure and tooling used for machine learning at a company. Keep in mind that these titles vary between companies.
You’re now a Machine Learning Engineer at Headspace! How did you land the job?
Headspace is the creator of sleep and meditation apps. I found this contract role at Headspace totally through networking. Mariam at FourthBrain told us to do as many informational interviews as possible, which I took to mean get in touch with people in the industry and become friends. Through a mutual contact in San Francisco, I connected with someone at Headspace and we hit it off — we even started a machine learning paper reading club together! When there was a job opening on their team, I applied for it. It wasn’t a good fit for me, but it got me in touch with the hiring manager. After the interview, Headspace contacted me with contract work. I could say I landed this job through networking, but the key was making friends with people that did what I wanted to do.
What kinds of projects are you working on at Headspace?
I’m on the machine learning team right now, and I’m getting to do real machine learning work. There are six of us on this team with two new teammates coming in soon. I’m getting ready to launch my first model, which is mine from the ground up. The best part of this contract position is getting to work with amazing, senior-level people on my team.
Are you using everything you learned at FourthBrain on the job?
Everything I learned at FourthBrain was relevant, and it factors into my decision-making on the job. If someone wants to know my opinion, I’m using the intuition and knowledge I developed at FourthBrain. In terms of boots-on-ground work, I’m working with some familiar things, but also learning so many new things.
You’re also part of DarkCirrus AI — tell us about that side project!
DarkCirrus is the startup of a FourthBrain peer. He’s doing awesome work in the Computer Vision field. I’m doing consulting on OCR/NLP work that adds to what he was already doing when we were at FourthBrain together.
How is your understanding of music helping you in your machine learning career?
There are two big ways that my music career is helping my new machine learning career:
Looking back on this experience, was FourthBrain worth it for you?
FourthBrain was totally worth it for me. Technically, I think anyone could do this on their own because the resources exist, but I would not have done it without FourthBrain because I struggle to impose structure on myself like this program did. The resources and people at FourthBrain are awesome. The experience of working with a cohort made the whole experience better.
Going to FourthBrain was worth it from a money perspective, too. For me as a musician, I’m making more money in my machine learning career. Plus, FourthBrain was also worth it from a resource perspective. The person who said you can learn anything on YouTube is correct, but that’s only true if you can manage that.
Flatiron School experts break down the differences between college and data science bootcamps!
A helpful guide for beginners with an expert from Tech Elevator!
How Doug launched a data career after Thinkful’s Data Analytics Bootcamp!