Cracking any interview requires preparation and in the case of data science it is not restricted to performing well on the big day alone. An aspiring data scientist is expected to prepare across multiple fronts. Here, we provide you with a insight into the levels of preparation required and how to go about it:
Preparing For The Various Rounds Of Interviews:
As the most basic round for any interview, it covers fundamental topics such as English language comprehension, quantitative aptitude and analytical reasoning. While this round requires minimal preparation, lazily scanning through your Wren and Martin and high school grade quantitative aptitude questions will certainly help in brushing up the important concepts.
Technical And Problem-Solving Interview Round
This is where your technical grasp over the subject is tested — especially your programming language proficiency, knowledge of statistics, optimisation and machine learning. The main languages that one is expected have mastery over are Python ,SQL, R, Scala and Tableau. Knowing Java and C++ also helps in adding depth to your programming skills. Since this round will deal with the brutal basics of the languages which you have stated in your CV, brushing up on their basics is of utmost importance.
Both freshers and experienced candidates will be questioned on key and rudimentary topics like:
- Probability — Random Variable, Bayes Theorem, Probability distribution
- Statistics — Sampling Theory, Hypothesis testing, Summary statistics
- Statistical models — Linear Regression, Non-parametric models, Time Series
- Machine Learning — Bayesian ML, SVM, Decision Tree, Logistic Regression
- Understanding of neural networks
The problem-solving round often involves a case study which requires the candidate to define the problem for the scenario presented, and explain the business impact of the solution.
During the process, bringing in examples of case studies and findings of research papers to support the solution will help improve the candidate’s score. The round may also require you to evaluate the robustness, scalability, implementation issues, and so on, of an existing plan and provide alternatives. Freshers may also be asked to explain the mandatory projects carried out for the fulfilment of their academic courses and the rationale behind the methodology and solutions. Hence, it is important to be well-versed with your projects.
In the case of more experienced candidates, they may be tested along similar lines. But they may be asked to talk about their real world projects, questioned on the related domains and the impacts of their model on the business.
Willingness To Learn
The debate surrounding “building talent vs buying talent” has employers split into two groups. On one hand, building talent helps employers incubate and nurture talent according to the needs of the company, while on the other hand, buying talent provides the incentive of hiring people with highly specialised skills. As a data science aspirant, it is important to showcase the skills you have already acquired and honed, but also express willingness to learn in the current job and being open to adaptation.
Showcasing Your Inner “Unicorn”
Roger Huang of Springboard says:
“A data scientist is a unicorn that bridges math, algorithms, experimental design, engineering chops, communication and management skills, but they aren’t specialists in every aspect.”
In the current business environment, a data scientist’s role is one of a bridge between multiple facets of a business. Though it is impossible to be an expert in all aspects of it, a data scientist has the unique distinction of being able to ideate and provide solutions across multiple disciplines. Thus, this point ties in with the previous one where it is important to make your technical individuality felt but in the process show the potential employer that you have what it takes to be the unicorn that a good data scientist is expected to be.
Before The Interview:
Build A Portfolio Of Projects And MOOCS
It is true that most employers do not look for a candidate with a doctorate in Data Science. But what they look for is a candidate with some hands-on experience in it. While it is easy for experienced candidates as they have a body of work to show, freshers too can independently create some of their own. One of the best way to do this is by carrying out data science projects. There are ample datasets made available specifically for this purpose on public domains such as Kaggle, GitHub, Google Cloud Services, Amazon Web Services and MNIST, among others. Carrying out projects and participating in open challenges convey your initiative in taking on challenges and working on solutions to tackle them.
MOOCS (Massive Open Online Courses) are another way of adding weight to your portfolio. There are many popular platforms such as Coursera, edX, and FutureLearn, among others, which provide various courses (most of them free) related to data science. Along with academic knowledge acquired during your graduation or post graduation, taking such courses provide exposure to different and focussed application of the knowledge. To a prospective employer, this presents the candidate as a well-rounded person with a participatory approach in the field.
Network With Peers: Stay Up-To-Date With Who’s Who, Follow Trends
Networking with experts from one’s field of interest is vital and data science is no exception. Following experts in the field of data science and data scientists on professional platforms such as LinkedIn will keep you up-to-date with the latest trends in the field. It also provides a great opportunity to learn more about their work and their take on other people’s work. A quick glance through your profile on these platforms (which most employers do these days) gives the employers a peek into your participation on social media.
For The Interview:
Learn About The Position That You Are Applying For
Yes, you are applying for the position of a data scientist. But what kind of a data scientist?
It is a very broad term and given the nature of the market demands, it is not a generic one anymore. Roles such as an analyst, market researcher, statistician, and project manager, could also be included in the definition. Different businesses such financial services, e-commerce, marketing and so on, employ data science; and depending on the nature of the business the required skill sets may differ. For example, the banking, financial services and insurance (BFSI) sector requires a stronger theoretical understanding and expertise, while e-commerce requires a more model-oriented expertise. Also, the nature of the business that the employer is engaging in demands one to have a basic understanding of it as the data solutions sought will be in the realm of the business. Hence, before applying for the position, it is important to know the specific demands of the job.
Additional Cosmetic and Brownie Points:
A Tight Resume
A candidates CV or resume proceeds him or her. The proverbial first impression really does apply. That is why it is important to present a good resume which is crisp, concise, and highlights important aspects such as an advanced degree in the subject if you have one, your experience with data (projects, hackathons, etc) and understanding of business domain, among others.
Go Through Questions Asked In Previous Interviews
Many companies make questions from previous interviews available on sites like Glassdoor for candidates’ reference. Going through these questions have two advantages. One, it provides the candidate a general idea of the nature of questions that may be posed. And two, mentioning your research for the interview when a familiar question is asked will definitely help score brownie points with the interviewer.
Concentrate On Your Effort, Not The Outcome:
It is impossible for all of us to succeed in all our ventures all the time. A job interview is no different. While some of us may ace an interview with utmost ease, some may fall short. In such a case, it can serve as a valuable source of feedback and aid in self-improvement which can in turn intensify one’s preparation and reinvigorate their efforts.
Try deep learning using MATLAB