Why is it hard to get a data science job?July 9, 2020 2020-12-03 21:19
Why is it hard to get a data science job?
Why is it hard to get a data science job?
Tons of proactive data enthusiasts spend hours, days, weeks and even years to learn data science. However, it just seems odd that their professional portfolio isn’t getting them a call back from the organization they wanted to work at. If you fall into the same category, this article is aimed at helping you understand why it is hard to get a data science job and how you can work on landing your first job in the field.
Realities of the data science job market
An aspiring data scientist is usually as optimistic as an entrepreneur is with his/her new startup idea. However, the market decides whether an idea succeeds or fails and the same goes with the data science job market.
Here are some of the things that you need to know before applying to a data science job.
1. The supply of personnel is more than the demand of organizations
This is the biggest obstacle that lies ahead of you. There aren’t just enough organizations posting data science vacancies in comparison to the people who are applying to these jobs. This means that the general population has caught onto the data craze way before the companies have and the supply is blowing out of proportion to the demand.
2. Startups tend to recruit data analysts rather than AI engineers
Do you wish to start building a production-ready neural network on your first day at work in a startup? Well, you should know that getting into a startup in a data science role does not necessarily mean that you will be building a neural network anytime soon. This turns to be an issue sooner or later for most Deep Learning fanatics.
It is important to understand that digital products are coming at a place where their systems are finally integrating CI/CD techniques into their workflow. Building Neural Networks at this stage is not only impractical for them but also not a good return on investment. Thus, it may be hard for you to gobble down this fact if you are unaware of it early on.
But, if you are eager to get your hands dirty early on, you can always aim at getting a data science job in a well-established company or an AI-based startup. However, if you stick long enough at a company, you’ll certainly be able to try different avenues on your own lead given that your motivation doesn’t go away by then.
3. Businesses prefer data science personnel who are comfortable with business lingos
Nice job learning all the classification and regression techniques but have you ever thought that you may now have to learn what KPIs, ROI, Breakeven point, etc. mean?
Knowledge of basic business lingos is becoming more and more necessary each passing day because businessmen want to hire people with good soft skills on top of their hard skills. This is because they see people in data science roles as creative and decisive individuals who can help one’s business make a good business decision using a data-driven approach. Here is a list of Business Jargons that are essential for data scientists.
Hope the heads up given above will be handy to you. Now, let’s move onto how you can actually get a data science job that you can love and commit to.
The guide to getting your first data science job
Sorry if the above issues demotivate you about your chances. However, a true entrepreneur never quits and thus, neither should you.
The following guide has been put out in a step-wise fashion for you to have an easier time reading.
Step 1: Follow all known data science company and their C-suite executives across all social platforms (especially LinkedIn)
Companies put out vacancies when they need someone to fill in a role that feels missing. Thus, they put out such call-outs across social media. If you have information that someone is hiring, you can learn more about them and write your cover letter like you had been waiting for that vacancy all your life. Quality over quantity.
LinkedIn is a special place for you to do this since C-suite executives are themselves sharing vacancies. However, you should apply for it through the mailing address rather than messaging them with your portfolio since that brings on a sign of desperation and we do not want to give that impression.
Step 2: Programming language selection – Learn R and Python in conjunction
Most companies prefer hiring aspirants who know Python but who have also tasted programming in R. This is good to have for companies since it shows that the aspirant is open to learning new languages and frameworks (such as Tensorflow or PyTorch) to get the job done.
Knowledge of R is good to have but knowledge of Python is a must to have in most cases. Furthermore, your portfolio should show that you are very comfortable working with the following programming libraries:
- Python: Numpy, Pandas, TensorFlow/PyTorch, Scikit-Learn, and Matplotlib.
- R: GGPlot
Step 3: Start playing with data as much as you read about it
Your work portfolio speaks more about you as an aspirant than any other medium. Thus, you should have links to published papers, research blogs or open-source data science projects at a bare minimum. This allows hirers to know your thought process, your tool of choice and your approach to solving problems.
To start off, pick a competition on Kaggle that seems interesting to you. Go to the public kernel section of the competition, find the highest-ranking kernel doing EDA (Exploratory Data Analysis), and go through the author’s thought process. Now, take a deep breath and work on the competition by yourself and share your kernel publicly to receive feedback from other data science enthusiasts competing in the platform. It usually takes about 4-5 competitions by the time you finally start doing note-worthy work. Here is an extensive hands-on course on Exploratory Data Analysis with Python for you to start your journey on working with data.
Your chances of landing your first job will certainly go up when you have proofread Kaggle kernel links on your portfolio that explains your thought process as well as domain knowledge. Bonus points if your kernels are from the same domain as the company you’ve applied to work for.
Step 4: Tailor your portfolio based on the job description and not based on the title
Job titles are usually kept to be more lucrative than what they really are. Read the job description carefully and understand what the role is about without looking at the title.
For example, a trend seen in the HR industry right now is that companies generally receive more applicants when they put out a vacancy for a data analyst rather than business intelligence analysts. Thus, a title may fool you into a job that you wouldn’t want to apply at or don’t have the knowledge for.
The list below shows you the actual title based on the job description keyword:
- Should be able to use SQL and handle ETL processes – Data Engineer
- Should be able to extract quality information from available data – Data Analyst/Business Intelligence Analyst
- Should know SQL and handle ETL processes, build robust AI systems and extract quality information from all kinds of data – (A three-in-one Data Scientist)
Step 5: Be patient and keep applying
The best things come to those who wait.
If you have put your time into doing your homework and sending tailored cover letters and a well proofread portfolio displaying your work, you should get a positive reply back soon. Wait and work on your skills in the meantime.
Step 6: Appear your interview as a human being and not as a robot
You’ve done it. You’ve received that good news of being shortlisted and now, only an interview (more than one in some cases) separates you and your new company from finally being together.
Interviews are generally the same across all fields of work and your confidence and knowledge are put at a test. If you think you can do it, you will. If you need help brushing up your data science skills before an interview, we have multiple courses for that on The Click Reader.
At the end of the day, getting hired is a game of both luck and hard work. Do your best and always have confidence in your abilities.