4 Ways to fail a Data scientist job interview
‘Data Scientist’ might well be the sexiest job of the century. But hiring one is anything but that. Actually, it can be excruciatingly painful for companies. It’s an equally big deal for aspirants to bag that perfect role in core data science, one which has a lot more to offer, than just a glorified title.
While machine learning is tough, training a human who can make machines learn can be tougher. One evolves through various incremental stages of expertise to become a productive data scientist.
For companies trying to identify one, it’s like finding a needle in the haystack. After years of hiring data scientists at Gramener, I’ve seen some conspicuously recurring patterns of skill gaps in the market. While there are hundreds of ways to fail an interview, these can be isolated into 4 broad paths.
The 4 pathways to Rejection
Given that only a handful from amongst thousands of applicants will crack that meaty machine learning position, it is helpful to understand where most people fail. For any aspiring data scientist or one looking to move up jobs, these are clear pitfalls to be avoided.
Becoming aware of one’s weakness is the sure and steady first step in fixing it.
Becoming a truly successful practitioner of data science involves picking up a specialized skillset. What better way to illustrate these role nuances rather than with a light-hearted analogy? We’ll compare this making of a data scientist with that of training to be a sniper, another cool job that calls for exceptional skills.
Let’s begin.. so, what are the 4 ways to fail a data scientist job interview?
1. Window-dressing the CV with machine learning buzzwords
As with any job, it may be tempting to tailor one’s resume by peppering it with industry jargon. And data science has no paucity of buzzwords. While this act of window-dressing does improve the chances of a CV getting picked by the automated scoring bots in HR, this can backfire rather quickly.
It’s not uncommon to find that advanced analytics skills claimed on paper actually translate to nothing more than basic familiarity with excel pivot tables, SQL queries, or Google analytics. Even if we set aside the time wasted, this poor tactic sets up candidates for big failure and a bigger demotivation.
For our aspiring sniper, this act equates to putting on the garbs of a soldier and picking up a gun, without investing the time needed in training to be one. As absurd as it sounds, there’s no fun for a sheep going in to hunt in a wolf’s clothing.
2. Reducing model-building to just making library calls
Many candidates who claim to know all about modeling, struggle greatly to explain beyond the model function calls and parameters. Even before asking what a technique like Random Forest does, a more important question is on why it is needed in the first place.
To be fair, a model is up and running with a single-line library call. But, machine learning is a lot more than that. One needs to understand, say, where logistic regression is more suitable than SVM. Or, when simple extrapolation is more powerful than forecasting techniques like ARIMA or Holt-Winters.
A good sniper needs to do a lot more than point and shoot. Actually, shooting is just 20% of the course in sniper school. One needs nuanced skills like patience, discipline and great observation to estimate target ranges from far.
3. Lacking the fundamentals essential for data analysis
While an intuitive understanding of machine learning techniques can serve as a strong plus for candidates, they often stop right there. Investing in hands-on training to master more fundamental skills like statistics and exploratory data analysis is often overlooked.
Modeling accounts for just a small portion of the analytics lifecycle. In any successful ML project, over half of the time is spent prior to that, in data preparation, wrangling, and approach. And almost a quarter of the time after, in model interpretation and recommendations.
Even as candidates flaunt 90% accuracy levels in projects, it’s a tragedy when they struggle to explain what a p-value is. It’s heartbreaking to see their diminished confidence in explaining why we need confidence intervals for models.
A firm grip on fundamentals is critical in all disciplines, and a sniper firstly needs to be a great infantryman. Of what use is excellent marksmanship, if one can’t fix a gun that jams or misfires in the midst of a battle?
4. Inability to apply analytics to solve business problems
It is clearly an uphill task mastering all aspects we’ve discussed so far. But we still miss a critical link in the chain, and this is where most interviews come to a screeching halt.
The ultimate mission for data scientists is to solve a business problem and not just analyze data or build a great model. This is the holy grail of data analytics. One needs to frame the right business questions, and evolve a sequence of steps to solve them. Even before loading any data into a tool.
When quizzed how a business can address their customer churn problem, it’s a conversation killer when candidates rush in with ideas of data analysis, or worse, toss around model names to predict churn. A better start is to probe on why customers sign up, the value they expect, and what influences business.
Imagine an expert sniper who knows it all, but can’t conceal and camouflage in the ground or pick out the right targets to eliminate. Such an individual is truly a dangerous person, and the bigger risk is internal for their own unit than for an enemy.
Wrap-up: In pursuit of data science
In summary, one must adopt a disciplined pursuit of data science by:
- Approaching a business problem by reframing the questions and evolving a sequence of steps to address the challenge,
- Applying the fundamental skills in statistics and exploratory data analysis to get a feel for the data and iterate the analytics approach,
- Choosing a selection of analytics techniques or machine learning models, and then engineering and interpreting the results for business users
- And showcasing these skills with the right positioning of one’s expertise, for an ideal role fitment
So, good luck bridging the gaps and to create a dent in the analytics job marketplace!
If you found this interesting, you will enjoy reading my recent article on how to acquire non-intuitive superpowers to rise faster in your data science career:
Passionate about data science? Feel free to add me on LinkedIn and subscribe to my Newsletter.
