Contact Info
198 West 21th Street, Suite 721
New York, NY 10010
youremail@yourdomain.com
+88 (0) 101 0000 000
Follow Us

Eleven most common objections in data science interviews (And how to handle them?)


1000x400 3

 

So, you are all set for your upcoming data science interview. You have probably double-checked on your knowledge and practical skills. You went through various checklists required for a technical interview. Now, you are counting on your experience to help you navigate through this job interview. So, it would make sense that you also go through a list of some common objections posed in a data science interview. Look no further as we have carefully forged a listing just for you.

Here is the list of eleven most crucial objections found in a data science interview, and how can you handle them!

Understanding these objections will not only help you dodge the ‘bullets,’ but also pave your way to succeed where many fail. Read on to ‘fill’ the gaps that may arise during your data science interview.

 

#1. Presenting a ‘monotonous’ portfolio

A ‘regular’ or ‘monotonous’ portfolio won’t get you killed! But it comes out as a common complaint by the data science recruiters. You should exploit this chance to showcase your organizational and communication abilities to the company. Why turn-off the hiring manager with the stuff ‘everyone’ has? Imagine the amount of effort you put in for the first-date.’  Then why not follow the same approach with your dream job opportunity? A thoughtfully crafted portfolio will reflect your detail-orientation, seriousness for the position and communication efficiency. The point here is to stand out from the crowd. Let your portfolio make an impression on the recruiter. Here is how you can create a meticulous portfolio;

  • Use the power of visualizations: After all, data scientists do this all the time. They use visuals to express data reports and so on. Displaying your projects, achievements, skillset, etc. in a visual format will serve you two-fold benefits. First, the interviewer can quickly & easily grasp what you have to offer to the company. Second, a carefully crafted visual portfolio will reflect organizational ability and creativity on your part.
  • Frame the information in such a manner that anyone reading your portfolio understands it in a matter of seconds, say, about 20 seconds.
  • Include these projects: Your portfolio should consist of data cleaning projects, data-storytelling projects, and end-to-end projects. All these projects involve algorithmic coding, data cleaning, and analysis skills. Such projects will straightaway let the interviewer award ‘points’ to you.

 

#2.  Delivering a ‘sloppy’ code

You can be certain that your coding and analytical skills will be put to the test in a data science interview. An algorithm coding test will directly show the technical value you will bring to the company. Writing a ‘sloppy’ code or the one with too many bugs might be the last thing you would want to do! A poorly written machine code is a commonly put objection by the recruiters. If asked to write a program, code an algorithm or work on a dataset, make sure you avoid these common mistakes;

  • While writing code, make sure you enter ‘spaces’ when required. Not only lack of spaces make the lines unreadable in code, but it also makes it hard to ‘edit’ or ‘correct’ if the program crashes or has bugs.
  • Write ‘only one statement in a line’ while coding. While this is not an error, but a great coding practice. While revisiting the code, you can read each line clearly and check for an error.
  • While cleaning a dataset, remember to remove the ‘outliers’ and ‘duplicates’ carefully. Also, make sure that you treat zero, specific values and null carefully. Not defining them correctly can lead to incorrect results.
  • While adding transformed or calculated variables to a dataset, do not assign numerical ids to the variables. Every company system has their way to generate ids. Assigning variable ids yourself can return erroneous results.
  • While before transforming the dataset at the end, make sure you give a thought to new possible variables. Usually, data cleaning takes up a lot of time, and the candidates get exhausted. Instead, start a fresh and think about potential transformation variables.

Following the above tips can significantly improve the quality of codes and data analysis model you build during the hiring process. Keep these in mind and deliver the best you can!

 

#3. Not being clear about the job role

This objection is self-justified. If you walk up to the interviewer and have no or vague idea about the job role and responsibilities, your competence and determination can be questionable! The field of data science has many closely associated job profiles. However, they differ vastly in the duties and skills required to undertake them. It is like taking the girl out on a movie and ‘not knowing her name and other basic details.’ Be ready to face her ‘cold-side’ then. While this analogy sums up the crux here, let us take a look at some pointers you can take care of!

 

  • Know ‘who’ they are looking for: An organization might be looking for a developer, data analyst or a data scientist. So, you should understand the fundamental responsibilities and skills required for each of them. It would make sense for you to apply for the position of data scientist if you know statistics and algorithms, and likewise. This cuts a lot of confusion for the hirer and the hiree.
  • Know ‘what’ they are looking for: Here comes the list of skills and knowledge background a company is looking for in the candidates for a particular position. For example, the position of a data scientist requires knowledge of languages like Python and R, analytical skills and experience with various database systems like SQL and NoSQL. While, a job of data architect mandates knowledge of Pig, Hive and Spark along with in-depth knowledge of ETL, database architecture, BI tools, data warehousing, etc. The gist is, make sure your skill set & experience matches the open data science position.
  • Know the company: Read the company’s website carefully. Know who they are and what they do. Read briefly about the founders as well. Understand what their products/services are, and where do they plan to go with the ‘big-data’ rush!

 

#4. No hands-on experience in data science or ML projects

This is one significant ‘red-flag’ objection by the companies while recruiting talent for data science profiles. A fair, if not rich, experience in data science and machine learning projects is highly preferred by the organizations. There is a high probability that the lack of such practical experience on real-world projects can keep an aspirant from succeeding in the data science interview. You have to make sure that you list them in your portfolio to reflect your technical competence. Here is what you need to take care of;

  • Include data-cleaning projects: A crucial must-have skill for any data science professional is to be able to clean the raw data and use it meaningfully. The interviewer will scan your portfolio for this project. You can even take-up such a project and assign it to yourself. It is all about the experience after all! Find datasets online and complete the ‘cleaning’ project using tools like ‘R Markdown’ and ‘Jupyter notebook.’
  • Include data-storytelling projects: These projects are majorly done for extracting valuable information from the data sets. In short, the data ‘tells’ the story about a particular problem, relation, and possible solution. For example, if you take a dataset for international flights and analyze the delays. A possible relation can be demonstrated between flight delays and loss to the airlines. A solution can consist of ‘optimizing’ the flight timings or routes. The idea here is to show the interviewer that you can deliver the business value to the company.
  • Include end-to-end data projects: Such data projects reflect your capacity to build user-grade operational models. The interviewer will award the highest points for such projects. If you hold experience, well and good. If not, start making an integrated, robust data project now! For example, you can build a project on ‘realty market’ prediction. You would require demographic and real-estate data. Then, you can build a model from scratch that ‘predicts real-estate market’ on the annual/semester basis. Of course, you are free to explore any other topic of interest.

 

#5. Lack of command over data science technicalities

Data science and data analytics are closely associated with machine learning and AI. It is indeed a field of high technical competence. It is obvious that a professional needs to be technically sound and abreast of the latest tools, technologies, and trends in this field. An interviewer can object if he/she senses gaps in your technical knowledge. Maybe you showed an excellent portfolio of fabulous projects. But, they can only be verified by testing your technical knowledge. Here is a checklist you should keep in mind before appearing for a data science interview.

  • Be thorough with the coding languages: As a data science professional, you should know the ins and outs of Java, C, Scala, Python, R, etc. Even if you have not worked with all of them, you should know the pros and cons of using these languages in the data science context. For example, some languages are suited for medium level data crunching, some for visualization and some are general all-rounders. Also, make sure you know how to code in at least two programming languages.
  • Be thorough with the algorithms: It is highly likely that you will be asked to write a code for an algorithm. It is recommended that you understand the standard algorithms like k-means, Apriori, AdaBoost, linear regression, logistics regression, etc. Polish your coding skills regularly.
  • Stay updated with the latest industry trends: Internet of things (IoT), machine learning and artificial intelligence, data analytics, and open source software are some of the newest data science industry trends to be named. As a professional in this field, staying up to date with these emerging technological trends comes highly recommended. The interviewer would be glad to know your knowledge and technical-opinion about this industry and its future.

 

#6. Lack of problem-solving and business acumen

Hiring managers can be stunned by your technical aptitude and skills. However, convincing them of your problem-solving abilities and business aptitude is a must. This is when they can indeed make a judgment about you. Often, aspirants display immaculate skills combined with in-depth knowledge. But, when it comes to solving real-world problems, they fail to deliver. Not only that, the companies want their data science professionals to carry business acumen as well. If you can understand a problem with a business point-of-view, you can deliver an equally apt solution. Afterall, big data is all about solving business problems, providing convenience to customers and refine industrial processes. Keep the following points in mind while sitting for a data science interview.

  • Offer your answers in a ‘solution format’ to the interviewer. Tell the interviewer how your knowledge and skills can be utilized by the company to achieve a feat. For example, you can suggest product/service improvements or a theoretical solution to a process the firm faces regularly.
  • Showcase your business aptitude by backing up your statements with a competitive strategy. If you have decided to talk about the company’s competitors, make sure you also talk about the measures your organization can take to outrun them.
  • Offer your solutions in the ‘business language’ and think like a business owner. Drive the interview conversation in a direction where you are comfortable giving a global outlook on big-data. Don’t hesitate to think regarding innovation and talk about it!

 

#7. Not being able to answer ‘basic questions’

You may fumble with speech if a beautiful woman starts talking to you, the same can happen when engaging in a face-to-face round with a data science professionals recruiter. We are sometimes unable to answer fundamental questions because of lack of presence of mind. Outside the interview room, we all know everything. But in there, we can sometimes act stupid and miss some data science question that we know very well. Here, your knowledge and experience are not in question. What matters is the confidence and presence of mind. We have compiled a list of some basic questions that may be shot at you during the face-to-face round. Take a look and expect similar questions in a data science interview.

  • Which language would you prefer for text analytics-R or Python?
  • Explain recommender systems?
  • Why is data cleaning a crucial step in data analysis?
  • From an unstructured data, how would you create a taxonomy to identify key sales trends?
  • How will you predict categorical responses?
  • When did you use the logistics regression function recently?
  • What is normal distribution?
  • List the key differences between univariate and multivariate analysis?
  • How will you select ‘k’ in k-means?

 

The list of such question is inexhaustible. You can refer this link to more such questions. The pro-tip is preparing well and deliver well!

 

#8. Wrong usage of jargons and terms

A seasoned or a passionate data science professional knows their field very well. It would be an assumption that you understand all the industry jargons and use them in the right context. Surprisingly, this comes as another common objection in the data science interviews. Candidates often misuse jargons or terms in the wrong context.

Like, calling a set of 150,000 observations ‘big-data’ or stating that a data scientist and data analyst are necessarily the same.

Such statements can be forgiven once or twice. Playing with them more than that will catch the interviewer’s ‘X-ray’ vision. If any ‘term’ or scenario catches you by surprise, the best policy is, to be honest. Don’t try to ‘play’ the recruiter. Be willing to accept and learn. You will be appreciated.

 

As a remedial measure, go through an updated glossary of terms used in data science and big data terms.

 

#9. Withholding personal views on the future of ML and AI

Data science walks hand-in-hand with machine learning and artificial intelligence. The interviewer would love to know your personal opinion on the future of this technology. But aspirants don’t usually converse openly during an interview. Hey people! Open up and put your opinion forward with some reasonable argument. Tell them ‘where do you see this technology heading to in the next 5 to 10 years.’ Tell them ‘what impact will it have on the industries’ and the ‘potential this technology holds within.’ Make such bold statements, and you will establish yourself as a curious and visionary professional.

 

#10. Being ‘overly’ creative

Hiring managers are always on a hunt for professionals that can ‘fit,’ ‘suit’ and ‘grow’ within their organization. Data science recruiters ‘fear’ two kinds of people. One who is ‘incompetent’ and second who are ‘overly creative.’ Creativity itself is not the ‘devil’ in this context. But highly creative people are often seen as ‘disruptors’ in an organization. Companies often find them to be a difficult ‘fit’ in their ecosystem (Unless it’s a creative job!). As a data scientist, you are expected to be organized, compliant and analytical. Being creative is alright, but make sure you project a balanced image (of yourself!) to the interviewer.

 

#11. Not sounding like a team player

Companies are run by teams. Individual talent plays a vital role in contributing towards a goal. But, big feats are accomplished by ‘collective talent.’ Remember, the interviewer will check on your team-spirit as well. So, you have to display individual efforts and collective team-efforts equally. In fact, put more weight on your team achievements. This comes highly recommended. In companies where emerging big-data technologies are being harnessed, millions of dollars are being poured in to create a quality workforce of data science professionals. Keep in mind to ‘list your collective achievements’ during a data science interview!

You just came across some critical objections found during a typical data science interview. Being aware of these ‘gaps’ will give you an edge over the others. Working on them will help you in the long-run as well.

 

Do you have something to ask or an interesting point to add to this list? Share with us in the comment box!

Post a Comment