Reflecting on my Metis Data Science Bootcamp Experience

I just graduated from the Spring 2016 NYC Metis Data Science Bootcamp (DS7 cohort) so I figured it would be a great opportunity to reflect on the experience. If you’re interested in learning more about Metis (i.e., you’re researching, applying, or preparing to attend), this post will provide you with valuable insider information about the bootcamp, including: the people, the curriculum, what it’s like attending class at the NoMad WeWork, the Metis resources and perks, my general likes / improvement recommendations for the bootcamp, and even some tips for future cohort members. I hope you enjoy reading!

Note: if you want to better understand my background / perspective or what Data Science is, please refer to my first post before reading the rest of this post.

The people of Metis

Instructors and support staff

The Metis teaching support system consisted mainly of two instructors and several TAs. From what I could tell, the Metis strategy is to pair instructors to achieve a balance of math and engineering expertise. I think this worked well because it provided us a more holistic learning experience and allowed us to “target” questions for specific teachers depending on the context. There was also an extended teaching support system that includes several TAs. They helped with everything from facilitating class logistics to answering our technical questions.

Our cohort had the interesting experience that one of the instructors had to leave a few weeks in due to a personal emergency. As a result, we had at least four other Metis Senior Data Scientists fill in the second instructor role for the remainder of the cohort. At first, the lack of continuity was somewhat difficult to get used to, but in the end I really appreciated the exposure to additional instructor perspectives.

Not to be forgotten, we also interacted with the Metis support staff which included program management, career support, and marketing. I found the Metis team to be great! Everyone from the instructors to the support staff was personable and really cared about providing us with a great overall experience.


One of my favorite parts of the bootcamp was working with and getting to know my classmates. Our cohort had 21 people with a wide spectrum of ages (early 20’s to 30’s) and industry experiences (just graduated undergrad, academia, tech, finance, and many others). IMO, the diversity of the cohort definitely inspired additional creative problem solving.

Our cohort was very friendly and social. Everyone was very supportive, always trying to help others or share best practices. We also found some time to socialize with each other. Almost every day there was a group of people that went out to lunch. There were also happy-hours, adventures to get ice cream, coffee breaks, Meetups, and much more.

Our DS7 cohort after our final presentations.
Our DS7 cohort after our final presentations.

In addition to our cohort, we had a good amount of exposure to the Metis alumni which I found very valuable. Alumni frequently hung out at Metis, working on Kaggle competitions, continuing their passion projects, or strategizing their job search. Having alumni accessible almost every day made it easier to brainstorm project ideas, troubleshoot problems. Alumni also gave us insight into what to expect not only during the bootcamp, but also after.

The Metis curriculum

Overall, the Metis curriculum aims for breadth of experience, covering a wide-range of data science topics in a short amount of time. During the 12 weeks, we focused on everything from math-heavy concepts (e.g., statistics, probability, linear algrebra), to computer programming concepts (e.g., Python, JavaScript, and SQL), to your more prototypical data science-y machine learning (supervised and unsupervised; regression and classification) and Natural Language Processing. The broad exposure allowed us to better understand the fundamental components of data science as well as which components we like / dislike.

Here are some of the projects I completed during the bootcamp to help you get a better sense of the tangible products that you may produce at Metis:

That’s the 10,000-foot view of what we covered in 12 weeks, but here’s what a typical day for our cohort looked like:

Time Activity
09:00 - 09:30 Arrive at WeWork, fill-up with coffee, get situated
09:30 - 10:30 Pair programming exercise
10:30 - 12:30 Lecture
12:30 - 02:00 Lunch
02:00 - 02:30 Investigation by someone in the cohort
02:30 - 06:00 Work on project

I feel like I would be remiss to talk about a typical day without mentioning the main class method of communication: Slack. The instructors and support staff mainly used Slack to communicate important announcements. We used it for more things than I can remember, but included sharing #resources, #pairprogramming, upcoming #events, and discussing #random nonsense. Overall, it was my first full-time experience using Slack. Compared to regular email, I really enjoyed the cohesiveness (e.g., easier to upload and find shared documents in one place), user-friendliness, and fun-factor (e.g., Giphy plugin).

What I listed above is a typical day, but the class structure was very flexible which I really appreciated. Sometimes I would go to the gym in the afternoon but come in earlier or stay later to make up the work. Not only did I do some of my best thinking during that break time at the gym, it was also nice to avoid the NYC gym crowds that hit after work.

One thing I was wondering before coming to Metis was what the workload is like outside of the 9 am - 6 pm class schedule. In talking to people from the prior cohort, it sounded like only a few people stayed much later than 6 pm every day. However, in our cohort there was usually a good group of people that stayed until at least 7:30 - 8 pm and sometimes even later. I know some people also came in on weekends to catch up on work. From my perspective, I did a good amount of work outside of class.

Attending class at the NoMad WeWork

Metis NYC holds class at the NoMad (North of Madison Square Park) WeWork. I really liked being located in NoMad because it is close to Madison Square Park which allowed us to hang out in the park on nice days - something especially nice during Spring in NYC. I also liked that there is a good amount of lunch options in NoMad, from Eataly and Shake Shack to Dig Inn and all the Korea Town restaurants. It was also walking distance from my apartment which was a huge plus!

Eating lunch with other classmates in Madison Square Park.
Eating lunch with other classmates in Madison Square Park.

If you’ve never worked in and/or heard of WeWork, I’m sure you’re wondering: what’s it like to work there?

Here’s what I liked about the NoMad WeWork:

  • Free food and drinks: Every Monday there’s some kind of breakfast treat at the entrance (e.g., they had a bloody mary bar on my first day). There’s also free coffee (with real mugs), fruit-infused water, and craft beer on tap 24/7.
  • Social events such as happy hours
  • Never a dull moment. We had class next to a modeling agency office so frequently they had models practicing their “cat-walk.”

Here’s what I think could be improved at the NoMad WeWork:

  • Keg stocking / tapping: a lot of times the kegs were out of beer completely or the taps were very foamy. I know, it’s “free beer” so I can’t complain too much…
  • Printing: they must really be trying to discourage people from printing because the system they have setup is a nightmare to use.
  • Noise: Metis is on the 3rd floor which I think is the most populated floor. As such, common-areas were frequently noisy. This didn’t bother me much, but I know it bothered others.
  • Slow elevator: the elevator at the 28th St. entrance takes FOREVER. I always used the Madison Avenue entrance which was much faster.

Metis resources and perks

Another thing I liked about the bootcamp was the excellent access to resources and perks.

We had access to several resources directly through Metis. For one, there was a recurring speaker series where data scientists from industry came to speak. Our cohort had speakers come from places like Google, Next Big Sound, and several other companies, but the lineup changes every cohort. In addition to the speaker series, there were several Metis-sponsored Meetups featuring some of the new and exciting work the Metis team members were doing. For example, I really enjoyed Mike Galvin’s talk on Word2Vec. As a secondary benefit, there was a lot of free food at the Meetups!

Outside of Metis, there are a lot of great tech / data science resources in NYC such as Meetups and conferences. My favorite Meetup experience was attending the NYC Data Engineering talk titled Google and Spotify Engineers on the Google Cloud Platform at the Spotify office. Below are some of the Meetup groups I attended, but this only scratches the surface of what’s available in NYC. I recommend you checkout to find ones that interest you!

Another NYC resource I enjoyed was the AWS pop-up loft in SoHo. I only went to get help once, but the space looked great and I know some of my classmates went to free classes there which they seemed to really get a lot out of. Definitely worth checking out!

Perk-wise, two of my favorites were for Amazon Web Services (AWS) and GitHub. For AWS, we were able to apply for $1000 in AWS credit through the AWS Activate program. The relatively hefty $1000 credit allowed me to do more experimentation with AWS such as using expensive EC2 GPU instances and hosting some of my projects for others to see. $1000 goes a long way with AWS so I’m looking forward to experimenting with even more of their services over the next few months.

For GitHub, we were given access to their Student Developer Pack. I just got access to it but I’m looking forward to using some of the resources such as the month of free access to any Udacity Nanodegree program.

General thoughts

I already shared many aspects of my Metis experience that I liked or could be improved. Here are some additional (more general) thoughts:


  • Having some structure to learning. Compared to self-teaching, the Metis setup allowed me to spend more time learning and less time trying to figure out what to learn and how to learn it.
  • Having on-demand support. Again, compared to self-teaching, it was really nice to have instructors available for support throughout the day. It was also valuable to be able to bounce ideas off the instructors and classmates.
  • How Metis encouraged us to write blogs. Writing my blog enabled me to build a more public brand for myself and helped me reinforce a lot of the things I learned during the bootcamp.

Room for improvement

  • Streamline the curriculum. Overall, I really liked the curriculum, but I think some things could be streamlined or prioritized, especially given the 12-week time-constraint. For example, less instruction on how to install software/tools from scratch such as Postgres and more emphasis on how to actually use the tools. Using an “as-a-Service” provider or Docker implementation for many tools would have saved some people a lot of time. I realize there’s value in getting people used to installing things from scratch, but I think there are a lot of new ways to do it more efficiently.
  • More hands-on instruction. Yes, we had projects, weekly challenges, and pair programming problems where we had to apply what was taught in lecture. However, I would like to have seen more hands-on instruction during lectures, such as interactive labs. I’m a very hands-on learner, so maybe that’s just me…
  • More exposure to A/B testing and recommendation engines. Having more exposure to A/B testing would be valuable since, from what I hear, A/B testing is a major real-world application of data science. As for recommendation engines, we had some informal exposure, and many people developed recommendation engines for their final projects, but a little more focus would be good IMO.
  • More “digestible” math content.

Tips for future cohorts

If you’re planning to attend a future Metis cohort, here are some tips I have for you:

  • Familiarize yourself with Python (especially Numpy and Pandas) before starting (or even before applying). One of the most useful things I did before the program was take this Udemy course: Learning Python for Data Analysis and Visualization.
  • Continually seek feedback. If I worked on a project for more than a few days without getting feedback from my instructors or classmates, I found myself going down an unproductive rabbit hole. Some of the most valuable conversations I had with the instructors were to check-in, discuss project next-steps, and get their feedback.
  • Prioritize your focus. We covered a ton of content which forced us to prioritize what we wanted to focus on, mainly through choosing our final project. For example, some people were more interested in recommendation systems so they chose to develop one for their final project. It’s also good to think about your goals for attending the bootcamp before you start so you can choose what to focus on accordingly.
  • Set a “fun goal.” Take a page out of Olympic gold-medal snowboarder Shaun White’s book and set a “fun goal” for yourself during the bootcamp. It doesn’t even have to be related to the bootcamp. For example, one of my “fun goals” was to find my favorite NYC bagel sandwich by trying at least one NYC bagel place each week!

My overall impression

Overall, I really enjoyed the Metis Data Science Bootcamp experience! I came away from the program with a more well-rounded skillset that gives me a starting point for solving data science problems and additional reach-back to a network of other data scientists. I highly recommend checking out Metis if you’re interested in data science!

If you have any questions, feel free to reach out to me via email at