Data scientists, AI practitioners, and machine learning engineers are among the most in-demand and highly paid jobs today. This has led to a surge in students and professionals taking up various courses in the past few years to enter this challenging field. However, one cannot become a data scientist or a machine learning expert overnight. It takes years of rigorous training and solid conceptual clarity to excel in this field, and only then can one call themselves a data scientist or an AI expert.
In a competitive job market like India, professionals seem to be looking for shortcuts to land coveted analytics and AI related jobs by taking courses and interview prep that “guarantee” they will crack the interview. Unfortunately, this mostly results in superficial knowledge that does more harm than good.
This fact recently gained a lot of attention when Nitin Aggarwal, Senior Technical Product Manager, Cloud AI Services at Google, highlighted it in a LinkedIn post, where he read in part, “Over the past few weeks, I've been on a whim to interview to build our team in India. I've met so many talented people and learned so much from their experiences. The common thread among the candidates who didn't get hired is that they were well prepared for the interview but lacked the 'real' work.
There were many candidates with years of experience. They had molded their previous non-AI/ML jobs into ML jobs and expected a senior level due to their years of experience. They had prepared a lot for the interview, attended courses and online prep classes. However, when we started discussing their projects in detail, problems started emerging. The answers were very shallow and textbook-like. They lacked a practical approach and it was easy to see that they came from some competition or exercise, not real life.”
Image: LinkedIn
Sprinkle words like AI and popular acronyms on your resume
Biswajit Biswas, Chief Data Scientist at Tata Elxsi, has an interesting take on why this is happening: He feels that by now almost everyone knows that the first stage of resume scanning is done by bots, and people understand how these bots work.
Biswas adds, “I agree with what Nitin says and have had similar experiences. There is a reason this happens. Finalists try to build their CV around AI and a few popular acronyms sprinkled in. So it makes sense that the CV will sound very AI-savvy, but the real truth emerges during discussions. This is even more the case at senior levels, where many are looking to take on the work of their teams (sometimes loosely connected) as their own, and many managers are not hands-on or involved in technical problem-solving. They are used to dealing with risk management and project management (which is also very necessary), but rarely get deep into code and technical issues. This is not surprising or new.”
Data science hype plays a big role
A lot of this has to do with the hype around the field of data science. Expectations of million-dollar salaries (don't take too much faith in media reports of entry-level employees landing such hefty salary packages), rapid career growth, and the opportunity to work with cutting-edge technologies right from the start of their careers can lead job seekers to project something they're not to recruiters. Venkat Raman, co-founder of Aryma Labs, points out why this is problematic:
- Many companies are in desperate need of top talent. The downside of experts labeling their non-AI/ML experience as AI/ML experience is that it becomes very difficult for companies to distinguish the good from the bad. The additional time that recruiters spend trying to figure this out is just too much.
- This practice doesn't reflect well on the applicants themselves, and Raman, who regularly hires data scientists and AI practitioners, says he has seen cases where companies have become upset by the practice and won't accept future job applications from such candidates.
Candidates and recruiters need to work together
This problem that the industry is currently facing cannot be solved by job seekers just interviewing. Both recruiters and job seekers need to work together to solve this recruitment problem. Biswas lists out some do's and don'ts that the industry must follow while recruiting for such positions.
- Candidates should focus on addressing what problem they solved and explaining “how” they solved it. It's simply a cliché to mention keywords, acronyms, or well-known frameworks.
- Similarly, recruiters should avoid entrusting the job of screening to bots or non-technical personnel. Screening and short first-level discussions to fine-tune candidates are helpful before moving on to more serious steps to secure interview slots for senior management roles.
- Recruiters need to broaden their scope and think longer term when considering the potential applicability of a profile.
“Shady” institutions also to blame
Apart from these suggested measures, Raman believes another problem of “shady” institutions also needs to be addressed.
He says, “Many aspiring data scientists refer to their training experience as an 'internship,' but in reality it is more like a cohort learning program without any industry oversight or mentorship. I would rather blame these shady institutions than the candidates, since it is these institutions that tell people to put 'Intern at XYZ' on their resumes and LinkedIn profiles. Industry leaders really should be calling out the issue of 'shady' institutions calling unsupervised cohort training programs as internships.”
A job in the analytics and AI field can be rewarding only if you have the right skill set. Otherwise, even if you manage to crack the interview, you will struggle in the role if the fundamentals are not clear. The analytics and AI dream becomes a reality if candidates focus on gaining real-world experience instead of superficial knowledge.
