Interviews

How to Get Your First Data Science Job: Interview with Michael Galarnyk

Knowing data science is great, but getting a job at it can be quite a challenge.

Today I have a special guest and he is going to reveal the secret you can use to get your first data science job.

Michael Galarnyk has much experience and he will show you proven techniques and strategies of getting data science jobs even when you are nobody in the industry.

Interview with Michael Galarnyk.

Interview with data scientist Michael Galarnyk. Michael Galarnyk will show you how to get your first job as a data scientist.

Photo of Michael Galarnyk

Godson: Good day, Michael Galarnyk, welcome to Cool Python Codes. Thanks for taking your precious time to be here.

Michael Galarnyk: Hey Godson, thanks for showing an interest in interviewing me!

Godson: Can you please kindly tell us about yourself like your full name, hobbies, nationality, education, and experience in data science?

Michael Galarnyk: My name is Michael Galarnyk. I am a marathon runner (Yes, I like to run 26.2 miles straight) and I am a Ukrainian American.

I’m currently finishing a part-time Masters in Data Science and Engineering at UC San Diego, and I have been working as a full-time data scientist for a company called Daymon Worldwide. Previously, I worked at Uptake and the Chicago Mercantile Exchange.

Godson: Which programming language do you use for data science and which is your favorite?

Michael Galarnyk: I use Python and R mostly. Python plays well with others and it just flat out fantastic to use.

Godson: What inspired you to become a data scientist?

Michael Galarnyk: I became a data scientist because when I was doing research on nanomotors, I realized that no matter what you do, it really helps to be able to analyze data. That and the job prospects weren’t great in NanoEngineering. One thing lead to another and I went back to school for a masters in data science.

Godson: Can you tell us about your blog?(What do you write about?)

Michael Galarnyk: My blog is about anything from installations (how to install Anaconda, RStudio, configuring AWS and such) to Python (Python basics, machine learning, PySpark, and web scraping)

Godson: Can you briefly tell me how can I become a data scientist?

Michael Galarnyk: That is a tough one. The easy answer is by grinding. It was and is a daily struggle. I would recommend:

  1. Make a GitHub and commit to it regularly (at least once a week). Basically, anything you do outside of work put on your GitHub. Having a portfolio can really help. It also helps you learn git better.
  2. Learn Python and take a lot of courses on Coursera or some similar online learning platform.
  3. Find something easy you want to code and code it (again put it on your GitHub). Iterate.
  4. Study machine learning algorithms.

See also: How to contribute to Open Source projects.

Godson: From your view, how long does it take to become a data scientist and what could be the challenges one has to overcome?

Michael Galarnyk:  You can get a job in about 3 months if you put in about 12 hours a day (maybe), but the learning never stops.

One of the biggest challenges is getting your foot in the door at any company. There are so many different ways to do it. It goes without saying that you have to study hard to be able to pass a lot of interviews, but sometimes getting those interviews can be nearly impossible.

Whether it be from networking, YouTube (seriously, it worked for me), blogging, or just flat out applying to hundreds of jobs online, it really helps to have many interviews to find out where you are strong and where you are weak.

Godson: Is there any educational requirement to become a data scientist? if there is, please enumerate them.

Michael Galarnyk:  Some jobs are strictly PhD. only especially for research like positions. A lot of positions ask for a minimum of a master and prefer a PhD. Some companies are okay with just a bachelor, but those jobs are rarer.

Godson: Is a data science certificate worth it when looking for a job as a data scientist?

Michael Galarnyk: Nope! They only help motivate someone to complete online courses which is valuable in itself.

Interviewers tend to care more that people know what they are doing, rather than a certificate.

Godson: Which IDE makes you more productive as a data scientist?

Michael Galarnyk: PyCharm

Godson: What data science tools do you make use of?

Michael Galarnyk: Python (PyCharm, Anaconda, PySpark),  R, AWS, and Tableau.

Godson: Which Python library for data science are you using all the time and why do you prefer to use it?

Michael Galarnyk: I love using pandas in conjunction with scikit-learn. Pandas really help with transforming your data, making features, and you can really use it for just about everything.

Godson: Can you recommend some good python books on data science?

Michael Galarnyk: I like the “Python Data Science Handbook” as it is 100% online. I like the step by step examples.

Godson: Which time do you like writing your code?

Michael Galarnyk: I like writing code at night and on weekends. I really do like to code.

If I am working on a blog post or YouTube video and I am almost done with it, I have no problem staying up the entire night to get it done.

Godson: The contents of your blog are very informative. I would like you to tell us, how do you come up with great posts?

What are your major sources of information for your great informative posts?

Michael Galarnyk: As far as my blog posts go, most of what I write are my works that I simplified for other data scientist to learn.

My most recent blog post (and the most popular to date): Using Scrapy to Build your own Dataset   is based on one of my current projects where I have to make my own dataset to determine what makes a successful crowdfunding campaign (I can’t say which website I scrape at work, but it is similar enough to the site in the blog post).

Michael Galarnyk’s YouTube video: Web Scraping using Scrapy.

Godson: One of the nightmares of any programmer is finding errors (bugs) in his program. Can you tell us your experience in debugging especially in data science?

Michael Galarnyk: Most of the time when I get a bug, I put that exact error into Google and it usually helps me solve my issue pretty fast.

I have found that the more I understand the various algorithms (KNN, Logistic Regression, and so on) and what sort of input they require, I can figure out why I am getting the error and fix it relatively fast.

Godson: How do you run your schedule?

How do you manage work, blogging, and school?

Michael Galarnyk: I am almost always doing something. Every day I try and spend at least 2 hours working on a blog post.

I have found that the knowledge gained from making a blog post helps me with being a better coder at work and makes me faster and more efficient at my homework.

Since I started my blog, it has even helped me get interviews and jobs. So many people reach out to me on LinkedIn asking me if I am looking for a job simply because they like my blog posts.

The beauty of a blog is that people at many big companies use Google to find answers to their questions. It just so happens that my blog appears on lots of Google searches now.

Godson: You said earlier that someone can get a data science job by networking, YouTube, and blogging. Can you tell us more about the networking strategy?

Michael Galarnyk: Honestly it was a very long-term strategy. You can’t build a YouTube channel or a blog overnight. Quite a few of my blog posts have gotten me interview requests by hiring managers at major social media companies, big 4 software companies, and tons of local companies in San Diego.

My pandas time series video got me a job. I have had the founder of Scrapy reach out to me over this YouTube video and corresponding blog post after it appeared on Hacker News. So many people have contacted me over my blog and YouTube channel.

It is ridiculous how much it has helped my career. Even something as non-jaw-dropping as Python Environment Management with Conda has gotten me a lot of attention and interest from companies.

People really forget that software engineers and data scientists often Google their issues. If these same people have their problems solved by reading your blog posts, they tend to think better of you.

The beauty of having a YouTube channel is you can have links to your blog posts in the comments section of your YouTube channel and get additional views to the blog posts. Over time, this makes your blog posts appear higher in Google searches and you get more interview requests.

Godson: How can someone prepare for a job interview as a data scientist?

Michael Galarnyk: Part of the problem with data science interviews is it can be impossible to predict what to expect especially at older companies that have no clue what they are looking for. I have had people interview me entirely on purely CS interview algorithm questions (bubble sort, linked list etc) which are poor predictors of success for data science jobs.

Some interviews are entirely based on how you would deal with class imbalance, what machine learning algorithms you would use for certain use cases, and SQL knowledge.

Godson: Which data scientists do you admire most and why do you admire him or her?

Michael Galarnyk: Kevin Markham as he is a fellow YouTuber. YouTube really helps with making a name for yourself.

Orysya Stus, She was featured in a really cool article by Microsoft recently.

Jake VanderPlas, he is the author of the Python Data Science Handbook.

Godson: Are you part of any Data science community?

Michael Galarnyk: I am not really part of any data science community outside of my fellow and former data science classmates from my masters at UCSD in data science. It is definitely something I need to work on though.

Godson: Can you share the tools you use to create your YouTube videos?

Michael Galarnyk:  Originally I used QuickTime (which is very silly). I detailed this approach here.

I also didn’t use a microphone (which I do now). Honestly, for YouTubers starting out, I would recommend using any means possible to get your initial videos online and then work on making your videos better.

Godson: How do you come up with the cost for your services for all the companies that contact you for a job?

Michael Galarnyk:  As far as cost/pay, you can only get paid what you can negotiate. If you are new to a field, you basically go for whoever will have you at whatever they pay.

However, once you get some experience, I would recommend you interview a lot (even if you have a job and are happy) and see what industry thinks you are worth. Once you know what you are worth, never take an offer under that amount.

Godson: What is your advice for anyone that what to follow your path?

Michael Galarnyk: Never stop learning. Always be working on building your knowledge and portfolio. Always be adding to your GitHub.

Godson: For someone that may need additional information from you, how can they contact you?

Michael Galarnyk: People are welcome to contact me through any of my YouTube videos, blog, Github or through twitter.

Logistic Regression using Python (Sklearn, NumPy, MNIST, Handwriting Recognition, Matplotlib) by Michael Galarnyk

Conclusion

I believe you enjoyed this interview with Michael Galarnyk. Indeed a lot of lessons can be tapped from this conversation.

As for me; I learned a great deal; one quality of Michael  I admire most is how he uses YouTube and his blog to get job opportunities.

Your Say

Let’s have the rest of the conversation via the comment section.

Should you have any contribution or question as regards this interview; kindly use the comment box below.

About the author

Rapture Godson

I am Godson; the brain behind Cool Python Codes. On this blog, I will make Python much fun and very practical. My up-to-date tutorials are based on my studies and they are very easy for you to understand.

2 Comments

  • Is data science all about python and R? Are they refer people using sql and excel to be data scientist s? What’s the difference between data scientist and data analyst?

    • Data Science is really about about Python, R and SQL. Excel is an excellent tool, but it is really limited. You can’t do machine learning in Excel for example.

      The difference between scientist and analyst is fuzzy. It depends on the organization.

Leave a Comment