The need for AI/ML is growing and more and more jobs are being created as data awareness is increasing and more data is being collected. However, hiring data scientists has not been an easy task – most of these roles are not yet filled.
On the other hand, data science is a very popular discipline. There are many fresh grads as well as folks who have taken various courses online and want to transition to data science. Finding a job has not been trivial.
Looks like there is a lot of demand and a lot of supply. But there is a skill gap.
What exactly is the skill gap ?
Lets see from the perspective of a hiring manager..
What are the main challenges when hiring for ML/data science roles ?
To decode this, we have had some interesting discussions with folks in the ML ecosystem on the gaps people see when hiring for data science positions. To keep the scope finite, when looking at beginner to mid level data scientists ( < 5 years experience overall),
The top gaps discussed so far can be broadly summarized as follows :
- Lack of problem solving skills in applying ML techniques to real products and making them work
- Lack of ML depth and breadth, though folks have worked on a few projects and write basic models
- Thinking from a customer/ business impact, and not just about model improvements
- Ability to dive deep into a problem, to figure out the real bottlenecks ML can solve
- Ability to deal with uncertainty and open ended problems as are common in real scenarios
- Humility to know what one does not know, and seek the right help and use available resources to solve problems the best way they can be solved.
- The desire and ability to stay with a problem long enough to make a meaningful impact
- Being receptive to adapting and evolving as the requirements and bottlenecks change
The best way to get these skills is to solve problems with ML for a real business with real customers for a while. Nothing beats the right experience, when it comes to building the problem solving muscles.
Interviewers often try to test these skills in some form in various interview rounds. This could range from asking deep concept questions to asking about the practical decisions one would take in scenario based rounds.
We at MachineLearningInterview.com are constantly thinking of ways help aspirants appreciate these aspects and help approach interview questions better from this perspective.
Do checkout our prep package, where we work on these aspects through mock interviews and other prep resources.