DS in the Real World

Nail Your Next Data Science Interview

Disclaimer: all opinions in this article are my own and not associated with any organization.

Over the last few weeks, I have been doing data science mock interviews for Grace Hopper Conference 2019 candidates. This is an effort organized by Khushali Desai with the goal to help prepare candidates who are interviewing at the conference. While mock interviewing, I am getting to learn about available data science resources, so I decided to share them here.

Data Science Resources

By now I have interviewed many candidates and some of them are really good and well-prepared. So, after each mock interview, I survey candidates on how they have prepared for the interview and resources they leveraged. These are some of the resources I have gathered:

• Kaggle for dataset & Kaggle for practicing on actual data science problems.
• Hackerank for coding (python, SQL)
• Theoretical: coursework at school/online course
• Analytics Vidhya for analytics questions.
• Google Top 10 Data Science interview questions and practice.
• Stackoverflow — an open community of coders.
• Glassdoor for understanding data science scope by company.
• Leetcode for algorithms (not required for DS)
• SQLzoo & Mode, W3School and Learning SQL book for SQL questions and practice.

Data Science Mock Interview

Before you participate in a real Data Science interview, make sure to prepare and practice. I have been using the below mock interview framework to help candidates prepare for the actual interview. You can use this framework to practice with a friend or a colleague.

Before I share, let me highlight that this is a rough interview framework for an entry level data science position interview. It by no means represents interviewing at a specific company but rather a general guideline for an entry-level data science interview. Data Science role vary company to company. Glassdoor is your best resources to find data scientist scope for a company.

Mock Interview Design:

This is the mock interview style designed for a 45-minute interview. You can think of this as a tech screen (first interview with the company) before onsite. I normally use the first 30 minutes for the interview and the last 15 minutes for feedback.

Interview frameworks:

1. Analytics/problem-solving.
2. Statistics
3. Machine Learning

Language: SQL, Python, R (could be any of the 3 or other if preferred)

Coding round: I use collabedit and create a link. I send the link to the interviewee before starting the interview.

Where to call? I use google hangout and enable my video so the other person can see me and feel comfortable.

Mock Interview Process: This is the process I follow for mock interviewing.

1. I will start by explaining the interview process. Ask them if they have any questions.
2. Introduce myself and then ask the interviewee to introduce themselves.
3. I like to do the coding exercise first so I can manage the rest of my time for other questions.
4. Most of my interview, I try to limit the interview to a total of 3 questions because I like to build on top of those questions to get the candidate’s depth of understanding in the subject area.
5. In the last 15 minutes, you should be providing feedback. Etc. when you explained x you didn’t follow the STAR method so I had to ask a lot of follow-up questions. You didn’t ask me questions before jumping into the problem. You forgot the answer and started apologizing, don’t apologize. Your application of supervised learning was incorrect because…
6. Thank them for their time and wish them all the best. You may tell them to reach out to you if they have further questions (not required).

Sample Data Science Interview Questions

Below are a few sample questions. The purpose of sharing these interview questions is to give you an idea on what to expect. As I previously mentioned, the data science role varies by company so make sure to do your research. The depth and complexity of questions vary by company, job level you are interviewing for, and the interviewer.

Opening Questions

1. Tell me about a data science project that you are proud of?

Analytics

1. Given this order table, build a sequential model that shows the growth of the business. Sales table [ORDERNUMBER, QUANTITYORDERED, PRICEEACH, ORDERLINENUMBER, SALES, ORDERDATE, STATUS, QTR_ID, MONTH_ID, COUNTRY_ID]. Define success metrics and then build sequential analysis using any coding languages.
2. I am a PM of UberEats and I am launching a new restaurant type offering, Seattle region. How do I find out the incremental lift of this change?
3. Follow to #2, design an experiment for testing this new feature. Identify and explain sampling techniques, launch criteria, measure treatment effect and analysis.

Statistics

1. When to remove outliers. Explain methods with use cases.
2. How would you explain the confidence interval to a business user?
3. What is the difference between type I and type II error?
4. This is a data frame of death in the military over the last 40 years. Perform and EDA on this data. Given the mean and standard deviation, explain what this tells you about the data?

Machine Learning

1. Explain Supervised and Unsupervised learning. When would you apply one over the other?
2. What is the difference between estimate and prediction? Which one is more robust?
3. Explain linear and logistic regression.
4. Explain multiple regression.
5. Can I perform a hypothesis test with regression? Explain how.
6. In a regression problem, what does the best fit mean? Explain.
7. You have a dataset with height and weight information. Identify response and explanatory in this dataset.

Final thoughts

Data Science is an umbrella term encompassing many fields within it. In many companies data scientists are generalists — read this blog to see what I mean by generalist — the above interview guidelines are for a generalist data scientist. I’ll say this again because it is important: make sure to do your research and fully understand the scope of data scientist role in the company you are interviewing for.

If you have questions, feedback, or additional data science questions that I should add to this blog, don’t hesitate to comment below.

Thanks for Reading!

If you like my work and want to support me…

1. The BEST way to support me is by following me on Medium here.
2. Follow me on Instagram here. I post lots of updates and interesting stuff here!
3. Also, be one of the FIRST to subscribe to my new YouTube channel here!
4. Follow me on LinkedIn here.
5. Check out my website, sundaskhalid.com.

📝 Read this story later in Journal.

👩‍💻 Wake up every Sunday morning to the week’s most noteworthy stories in Tech waiting in your inbox. Read the Noteworthy in Tech newsletter.