Data Science and the boom:
Data has always been the primary source of market information and performance. Even though the plain data does not necessarily portray the most accurate narrative, it is still the only parameter that gets anywhere near to it. Data Science is a lucrative, highly paid, and booming field in the world right now.
The US Bureau of Labor Statistics predicts a 28% rise in the number of data science jobs by 2026 i.e. roughly 11.5 million jobs in the USA alone. In India, a 45% increase in Data Science jobs has been predicted in 2021 taking the number of job openings to a whopping 16 million jobs!
Ask Henry highlights some staggering numbers in his comprehensive blog.
So, if you are eyeing a job in this field, the time is now.
Recently, we reached out to Charudatta Manwatkar, former Data Scientist at Curl to learn more about the Data Science industry and the interview process for these roles, and here are a few snippets from it……….
The Role of a Data Scientist?
Charudatta explains that a Data Scientist closely works with the people in the upper echelon of the hierarchy of a company to understand their goals and problems and determine how data can be used to achieve said goals and solve problems.
Most of the data in this modern world are unstructured and in the petabyte size range i.e. Big Data. This type of data comprises of the 5Vs- Variety, Volume, Value, Veracity, and Velocity.
To handle such large sets of data, data scientists take the help of programming and machine learning to arrive at results much faster and efficiently. The entire process of gathering and processing data involves:
- Initiating the discovery process by asking the right questions to the stakeholders and identifying their goals and problems
- Acquiring data- with ETL tools
- Processing, cleaning, and storing data
- Preparing an initial analysis of the data
- Prepare different models and algorithms
- Making use of machine learning, statistical models, and artificial intelligence (Ai tools) to arrive at a result
- Verifying the accuracy of the result
- Present it to company stakeholders with data visualization tools like Tableau
What are the required skills to apply for your first data science role?
Charudatta explains that the prerequisites for jobs in Data Science can get quite challenging as time passes on. As of now, candidates aspiring for a role in Data Science need to develop:
- Proficiency in Programming- Python. SQL, Java, R, Scala, and MATLAB
- Machine Learning- Natural Language Processing (NLP), Classification, Deep learning, and Clustering of data
- Data Visualization- Tableau, SAS, Python, R Libraries
- Big data- MongoDB, Oracle, Azure (by Microsoft), Cloudera
- Thorough knowledge of Mathematics and Statistics- Foundational knowledge in Linear Algebra and Calculus should do but statistics is a must
- Risk Analysis and research-Data mining, cleaning, and structuring
- Proficiency in effective communication (soft skill)
The Data Scientist Interview Process:
Charudatta explains that when he first joined Curl as a Junior Data Scientist, the interview process was not as streamlined as it is in 2021. However, exponential growth that has been observed in this field in the last 3 years has changed the way recruitment in this industry works now. Working data scientists contribute their time to oversee the recruitment. Here are some of the key takeaways:
The Resume sorting/ prospecting round:
All the resumes are gathered in a common portal from job postings and referrals by employees at the company. This portal is accessible to all the different teams of Data engineers, scientists, Business Intelligence Analysts, Data architects, Drone Analytics, etc.
The first resume round of resume screening is done by the HRs who sort out resumes based on the required role and prerequisites. However, since HRs lack the technical know-how, the team of data scientists is consulted, before arriving at a decision.
Members from each team use the portal to mark the candidates they like, based on their skills and interests and then reach out to them for a telephonic interview.
What makes a resume stand out?
Charudatta elaborates that in the world of Data Science hiring, every scientist has his/her own outlook as to what constitutes a good resume. However, there are a few markers that are absolutely essential to get noticed in this sector.
- Resume Length- a clean 1-page resume should be enough. Make sure it is concise, all the skills and projects are specifically highlighted.
- Relevant work experience- Make sure the job profile you are applying for, matches your work experience or projects. It becomes extremely important to prioritize your information to land your dream job. Focus on hard skills instead of wasting space on soft skills (which can find space in your cover letter). Having a non-trivial project is essential.
- Have a digital presence- Most data scientists research about a potential employee through their LinkedIn or GitHub profile. So, having a robust social presence on such professional platforms and forums like Quora might help you stand out in this digital revolution.
- Extra information- Make sure you highlight your performance in open source projects, community involvements, and data science hackathons. Another thing you can do is, highlight your efficiency in completing projects by markers like- an increase in revenue, high return on investment (ROI), etc.
- The number of referrals or validations- You should keep learning new programming languages and tools. If these skills are backed by certifications from accredited institutions, that helps as well. You can also ask your colleagues, mentors, supervisors to recommend/validate your skills on LinkedIn. Unlike what many people would tell you, each referral helps.
1st Round- The Telephonic Interview:
Telephonic interviews involve simple questions regarding major data science buzzwords and technical terms and systems. They will delve into your project experience to ascertain whether you will be a good fit for the company or not.
Make sure you research the job profile and the company meticulously and keep asking questions to demonstrate that you researched the company well.
2nd Round- Written Assessment:
In a 4-round interview process, the written assessment forms the second round. It is usually in the form of a home assignment. This varies from position to position. You will be assigned datasets to analyze or a coding assignment. Every company has its own policy on whether this round remains un-monitored or you have to log in with an interviewer who will monitor you while you do the assignment.
Make sure you use detailed visuals and showcase your deep technical knowledge so that you can make the assignment interpretable to the interviewer. This will also ensure the interviewer that your work will drive business value for the company.
You can try these practice tests organized by Testdome to refine your skills.
3rd Round- Interview with Data Science Manager:
The first part of this round is all about what you have done so far in your career. You will get questions on your previous job role, your projects, and your general know-how of the industry. You will also be asked about your general approach to datasets and analysis.
You will have to explain your projects, the methodologies you used, and the time you spent in seeing it through to the end. If you show signs of promise and knowledge then the interviewer will shift to a questioning line that involves aspects of what the company wants from you.
Note that sometimes, you may be asked to present some of your work or the written assessment to show your system of working with datasets and your proficiency in data visualization tools will be tested.
Here is a compilation of the top 50 questions asked in this round collected by the team at Simplilearn.
4th round- Interview with the Management:
This may usually be conducted by the Chief HR or even the CEO of the company as they usually want to be a part of the hiring process. This interview is seldom about asking you technical questions but simply to assess your general aptitude for different data science components.
This round is your chance to showcase your passion for the company so do not hesitate to ask questions about the role, about the workflow, about the different projects. In the end, mention how you would contribute to all those processes and why they should hire you.
Common Mistakes made by candidates and how they can avoid them:
Lack of effective communication:
Most candidates have the skills but lack the communication skills to present themselves to the interviewer. If you are not able to speak freely, you won’t be able to communicate your thought process behind your written assignment or explain your skills and projects in an impressionable way. Keeping a conversational tone throughout an interview round is very important if you want to create a lasting impression on your interviewer.
Much like in all technical fields, the lack of conciseness of the resume can be a cause for concern. Many candidates try to put everything that they have done in their resumes. This increases the number of pages to 3 or 4. In this virtual age and, after the pandemic, each and every job posting receives thousands of applications which is not ideal.
To stand out, you simply need to make a clean 1-page resume listing everything that is important leaving everything else for the interview rounds.
Lack of questioning:
As data scientists, you are expected to have doubts about almost everything. This “everything” includes your interview rounds as well. During questioning, many candidates forget to question the interviewer about the company’s expectations of him/her. Also, due to a lack of questioning, the candidates remain confused about their assessments and what is actually wanted of them.
Here’s an excerpt from Matt Przybyla’s article on 5 Data Science Interview Mistakes I’ve made:
“Not asking enough questions shows a few things:
- you are not interested in the company
- you did not pay attention well enough to come up with a question
- could show you are overconfident
- could show you are hard to work with”
So, make sure you ask questions, good questions. They will not only imply that you are in fact, interested in this opportunity but, showcase your passion for a role in data science itself.
Lack of clarity:
Many candidates primarily focus on getting an answer by using tools and projections rather than delving into the approach to getting that answer or interpret what it means.
This is because they either have gaps in their foundational knowledge or are looking for an easy way out. This may help them clear an interview but will definitely cause problems in the long run when business decisions and goals will drive what they need to work on.
How to Prepare for Data Science interviews?
Charudatta suggests that every aspiring data scientist should make learning from failures, a habit. The best way to prepare for data science is to get lots of practice with mock interviews taken by experienced professionals in this field.
“Another way to prepare yourself for your dream data science job is to regularly apply for fresher or entry-level jobs and internships. In applying for these jobs, you get to sit for actual interviews. If you fail, you try again and again till you get it right.”- Charudatta
Also, keep learning new things and stay updated with the latest trends in data science by engaging in new relevant courses where you can upskill yourself.
General interview tips for aspiring data scientists:
Always have backup devices (laptops and phones) for virtual interviews in case things go wrong. It does not help when the first few minutes of an interview are spent resolving technical issues. It helps to be prepared :)
- Dress professionally and have a proper background for your virtual interview.
- Be prepared to present by keeping all of your files ready to be showcased
- Have realistic expectations about data science jobs.
- Keep refining your skills and work on keeping your fundamentals clear
- Make your failures count.
Get to know Charudatta Manwatkar:
Charudatta worked as a Junior Data Scientist at Curl till the end of 2020 where he was involved in multiple projects. He was part of a team that made software that detected potholes and other damages to the road, from video and LiDAR data, Geo-tagged them using GPS, and calculated a rough estimate of the repair costs.
He gained experience and an interest in Natural Language Processing (NLP), he worked on building a corporate KYC tool that extracted information from noisy and semi-structured PDF documents for a global-level investment bank. He has 5 NLP specializations from Deeplearning.ai to go with his experience in the same.
After graduating with an integrated master's in Physics from BITS Pilani he worked in the Quantum Optics Lab at the Raman Research Institute, Bengaluru. He has always been keen to mentor young data scientists to land their dream jobs and is an avid writer in this space as well.
Check out the video highlights of our discussion below..