| | Michael Burnam-Fink | I've attended a lot of highly ranked educational institutions, including elite private colleges and the most innovative university in the United States. I have a doctorate, I've worked as a professor, and I've helped design curriculum. So when I say that Metis is the most focused, most professional, and most valuable educational program I’ve ever been a part of, it means a lot. I went from adjuncting at poverty wages in an academic field with no future, to data scientist at a financial institution, with a salary that I consider quite generous.
Metis is a 12 week bootcamp, that is truly 9-5 and then some. Each day at Metis begins with pair programming, where you’re given a problem and have to work out a solution with a classmate in about an hour. Pair programming solidifies the concepts that you’ve been learning, and is great practice for technical interviews and the real world. The problems are like those that you’ll see at a technical interview, and you have work with and explain your thinking to, people with a wide variety of backgrounds and levels of technical expertise. Sometimes you’re working with someone with a software engineering background, and you can whip up some kind of nifty recursive solution in minutes. And sometimes you’re the person who has a solution, and you have to help a classmate who’s struggling figure it out.
Mornings at Metis consist of structured instruction; lectures and prepared Jupyter notebooks that demonstrate concepts and best practices in data science. The coursework is brisk, and designed to take you from “What is a list comprehension?” to fancy stuff like implementing deep neural networks on AWS cloud servers. Lectures cover regression, classification, data structures, time series, natural language processing, and neural networks. Afternoons have talks about the job application process, and guest lectures from industry data scientists, as well as short presentations from students on things they’re experts in; everything from A/B testing to getting the most out of the Linux command line.
The meat of Metis are the projects. The first project is assigned, with an imaginary client and a small group, but the four remaining projects are up to you, with only the loosest guidelines, like “use linear regression” or “do unsupervised learning”. The best way to learn data science is to do it, and in projects you’ll learn problem formulation, data collection via web scraping, data cleaning and management, algorithms, and communication, as every project is finished with a four minute presentation to the class. The last presentation is in front of hiring partners at career day. The Metis projects are a great way to build a unique portfolio as a beginning data scientist.
Even though you’re working independently, you won’t be alone. The instructors, I had two in my 22 person cohort, are on hand to answer questions and help you think through your problem, and there are TAs, graduates of the last cohort, to help with debugging. Don’t get too attached to the TAs; they have a tendency to get hired away mid-course.
The bootcamp doesn’t require a lot of mathematical and programming proficiency when you begin. I recommend that you focus on the data science itself, and try and come in as well prepared as possible. Knowing your way around Jupyter notebooks and the Pandas and Numpy libraries will save precious time during the course. I really like Sweigart’s “Automate the Boring Stuff with Python” as as a place to start if you’ve never coded before. For machine learning, Raschka’s “Python Machine Learning” is a great desktop reference; I think half the class had a copy by the end.
I am very satisfied with Metis. The program, staff, other students, and ongoing career support were all top-notch. The only thing which I thought could be improved was the data visualization portion, which was merely serviceable, but I hear that it is being updated for 2019 to take into account Bokeh, Tableau, and Plotly. Obviously, you’ll get out of Metis what you put into it, but if you’re willing to work hard, it is an incredible opportunity. |
