Interviewing Tips for Data Science Hiring Managers

Based on my experience on what worked for me.

Few things

– Ask the person to describe how they would go about dealing with a Project that I am familiar with. Data pipeline, Model Building, Productionalizing

– that would be helpful to me how the person skills, experience and attitude would match to my needs and expectations

– since it is a very broad questions, follow up questions can be asked on choices made, strategy to deal with challenges, working in team , reporting status, managing client expectations and relationships

– I think it is very important to encourage the person to be able to express him/herself as they have no idea what the interview was going to entail and they were not preparing all their life to fill the particular requirement I have.

Case study

Subject: Me

Machine Learning Scorecard – 15 Jan, 2020

Statistics – C

Linear Algebra – C

Regression – B

Classification – B

Support Vector Machines – C

Decision Trees – B

Dimensionality Reduction – C

Deep Leaning -P

CNN – C-

Recurrent Networks -F

Representational Networks -F

GAN – C-

NLP – C-

Computer Vision – C-

Kaggle Courses – B

Kaggle Competition Win – F

LinkedIn Learning – A

Debate with LinkedIn experts – F

Legend

A – Build end to end model, debate it with an expert, be an expert of latest popular model even if it came to prominence yesterday, all this without any reference

P – Same with as many references as possible

B, C – somewhere in between

A+ – do something innovative, win a competition

F – yet to start seriously

Learning platforms

Completed courses – 88 ( includes some learning paths, actual course count > 100)

In Progress – 1,022

Other platforms

Udemy

O’Reilly Media

kaggle.com

fast.ai

Youtube

MIT OpenCourseWare

My basic argument is that even with P I can build a worthwhile model. However, everybody is expected to be A+.

Even if your model may not win a competition it is better than status quo if it is a greenfield implementation.

In most circumstances, the new person coming in, hasn’t had the same experiences that your existing team has had. If you have an existing team, you would be aware that it took time to gel. It is well documented in HR management literature – the stages all teams go through.

I have worked for close to 30 years and worked with and met lot of technology and business professionals. A large majority were/are not A+.

Yes, I agree it is better to err on the side of caution but it may be worth a leap of faith based on common sense developed from experience.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.