Insight data science interview preparation

• new project in kaggle (using codes and new techniques)
• crack on data scientist interviews (basic statistical questions..)
• behavior
• practice on my background

Application

• I applied online. The process took 2 weeks. I interviewed at Insight Data Science (New York, NY) in October 2017.
• I applied online. The process took 3 weeks. I interviewed at Insight Data Science in November 2017.

Interview

• Apply online: Fairly thorough application (describe background, list skills, eligibility, describe data science projects). After successful online application, additional interview is in two steps and typically done remotely: 1) “Phone” screen interview - “Phone screen” is really an online interview (e.g., using zoom) with data science program director (one of two) and program manager (somewhat HR). 2) Technical interview - Completed through a screen share with data science program director.

• online application followed by two rounds of skype interviews. first round was to know your background and your interests in transitioning to industry. second round was presenting a technical demo. interviewer was very nice. the whole process was pleasant.

• 30 minute interview with one of the program directors for the Silicon Valley Session. I applied during the early deadline and got an interview within 2 weeks of submitting the application. Some aspects of the interview: Describe your research. Do a short demo of an analysis you did. Relate it to business in some way. Then found out I got an offer 2 weeks later

• Applied to the fellowship program and made it to the interview stage. Did a video interview which required a coding demonstration. In preparation for this demonstration, I learned to code some machine learning models in Python. I was able to explain my code to the Interviewer without much difficulty, but with a little self-deprecation on my part. Gave an idea for a Fellowship project that I could complete that the interviewer said was a good idea. Overall, felt good about the how the interview went although I didn’t vibe too well with the interviewer.

• 1st round: the director introduced her background and the mechanism of Insight program, followed by some general data-science-related questions and my research background. Then I spent ~10 min presenting a demo of my data-science-related project. 2nd round: another director chatted with me about my research background. There were some technical questions regarding my data-science-related projects, but generally very easy. Both directors are very nice.

• First applied online, pretty straightforward. Answered a few question about skills and previous experiences, and they only wanted two-three sentences per question, which I appreciate. Then went through two rounds of phone interviews. Round 1 was more or less reiteration of application questions - why data science, what kind of company, etc etc. Round 2 was technical, was asked to do a demo with a short piece of code.

• Video interview with a Program Director. Asked for a description of dissertation work, then an analysis demo showcasing the types of data you work with, and what your plans are for the future.

• I interviewed at Insight Health Data Science (Boston, MA). Thirty minutes to one hour Skype interview with one program director (may from other programs) talking about the application. They asked questions like “show some code for your current project/side project”, “what kind of project you would like to work on at Insight.”

• I gave a brief demonstration of some code I wrote for a past project, and we discussed my background and career goals.

• I applied online which involved submitting a CV and answering a few short questions. The bulk of the interview process consisted of a screen sharing session where you had to share a data science/ analysis project you had worked on. This was followed by questions about what you would do to answer certain questions about the data, and what other algorithms you could have used. The most important aspect was to convey that you were naturally curious, a fast learning and enjoy diving into large data sets.

• Do a data science project and demo it live during the interview. The interviewer seemed very interested on the work I did and was eager to know more, asking very interesting questions.

• I was asked to have an example of code prepared. We spent a good part of the interview going over that code.

Interview Questions

Phone Screen - questions about background, why leaving academia, eligibility (i.e., ability to stay in location up to four months), what I might take on as a project.

Technical interview - share screen and display a project of interest (can range from for fun project to kaggle competition idea is to demonstrate as much data science as much as possible). Take through explaining common data science techniques (i.e., cleaning a dataset, feature engineering, feature selection, modelling, visualizations).

Advice coding doesn’t need to be OOP but should be decent enough not hacked together (i.e., some demonstration of control structures should suffice).

• Tell us about yourself
• Why I want to move from academia to industry (I have a PhD). Why do you want to transit from academia to data science?
• Why do you want to be data scientist?
• Why insight?
• Describe your PhD research

• What companies are you interesting working for.

• What is an idea for a project you would like to work on at Insight. What project do you think you will be doing at Insight?
• Walk me through some of your code.
• you demo interesting analysis work you have done, preferably different from your PhD research
• The interview goal was to see if you would be a good match and if you already had enough analytical skills to survive the program
• Describe a project that you worked on that involves data analysis and coding?

• Technical interview - share screen and display a project of interest (can range from for fun project to kaggle competition idea is to demonstrate as much data science as much as possible). Take through explaining common data science techniques (i.e., cleaning a dataset, feature engineering, feature selection, modelling, visualizations).

• The interview consisted of behavioral questions and a data analysis demo (with screen sharing)
• After filling out the online application form, the second round consists of a 30 minutes skype screen share session where you are asked to show one of your previous coding or data science projects. They will ask questions about this and also other general questions about your motivation about data science in general and the program specifically.
• why did you choose one algorithm over another
• Have you done anything that has been related to business?
• What kinds of statistical models have you used in your research? Can you explain one to me?
• What project do you want to do as part of your Insight fellowship?
• Why do you want to work in data science domain?
• Which specific space you would like to participate?
• What’s your research background?
• If you have all the data in the world, what would you do?
• Where do you see yourself in a year?
• Show some code for your current project/side project
• What kind of project you would like to work on at Insight.
• I was asked to explain my rationale in selecting a particular statistical modeling tool I used in the project I demonstrated.
• I was given a large quantity of yelp data and asked to find any interesting patterns
• Why do you want to transition from Academia to data science?
• What motivates you? growth, fun, rewards (known popular research)
• why would you choose analysis technique X over technique Y? (rationale and trade-offs)

• how to deal with spurious or errorneous data?

• Do you have experience with longitudinal data analysis? What Are Longitudinal Data? Longitudinal data, sometimes referred to as panel data, track the same sample at different points in time. The sample can consist of individuals, households, establishments, and so on. In contrast, repeated cross-sectional data, which also provides long-term data, gives the same survey to different samples over time.

Longitudinal data have a number of advantages over repeated cross-sectional data. Longitudinal data allow for the measurement of within-sample change over time, enable the measurement of the duration of events, and record the timing of various events. For example, suppose the unemployment rate remained high for a long period of time. One can use longitudinal data to see if the same group of individuals stays unemployed over the entire period or if different groups of individuals move in and out of unemployment over the time period.

The NLS surveys gather detailed longitudinal information about the lives of six groups of men and women over time. Each survey group (cohort) consists of 5,000 or more original members. Each cohort was selected to be representative of all people living in the United States at the time of the initial interview and born during a given period.