Job Interviews: A necessary evil

An interview can be defined as a formal conversation conducted to assess the suitability of a candidate for a specific purpose. In this article, I am going to focus solely on job interviews, although some of the points discussed may also apply to other types of interviews. Most job selection processes involve several rounds of interviews intended to evaluate the candidate’s qualifications. However, based on my experience, I have found the opposite to be true. I personally view interviews as a rejection process rather than a selection process. I have previously written about my tragic interview experiences when I was seeking employment after completing my master’s degree. Now, it seems I have lost faith in the process. So why do I still consider it necessary? Well, we need some form of filtering process as not all applicants may be suitable for the position, and there are usually fewer positions available than the number of applicants. Despite this, I despise interviews for many reasons.

You might have heard that success is a journey, not a destination. What about careers? Aren't they journeys too? Careers are not one-time stints where you clear the interview and are set for life, except perhaps in some government jobs. One must continually perform and get things done, yet I have never seen any interview round that ensures this. If a career is a race, it is more akin to a marathon than a sprint, and you would not judge a marathon runner by their sprint performance. The two require different skill sets and have more differences than similarities. Let me provide a real-life example to explain why I use this analogy. I am acquainted with someone employed at MAANG, who are known for their rigorous interviews; yet, this individual’s programming skills are lacking. This individual still uses reserved keywords for variable names, which can cause conflicts in the code and lead to errors. Although their code may run for a while, it ultimately fails in the long run due to these issues.

An interview is meant to evaluate a candidate for a particular position, but we often overlook an important aspect: it is relative to the interviewer(s). Most interviews are not absolute; they are significantly influenced by the interviewer(s). As an interviewee, I have given incorrect answers, but more often than not, I have been rejected due to the interviewers’ shortcomings. An interview depends as much on the interviewee as it does on the interviewer. The metric is not solely the projection of the candidate onto the interviewer, but also that of the interviewer onto the candidate. The dot product is symmetric, after all. Rejections hurt, but it’s even more painful when rejected for having knowledge or expertise.

With the AI revolution, most tasks can be automated, including coding. With advancements in AI, such as large language models, individuals can easily generate SQL queries for their use cases, yet there are interviews that evaluate certain skills. For instance, I still fail to see how being asked to code a circular queue during an interview would help evaluate my skills as a data scientist. If faced with any such requirement, I can always refer to online resources or use a search engine like Google. Then, one might ask, if everything is already available, why would one hire me? First of all, knowing what to look for is one of the primary skills, and coding is not software engineering. Companies need people to solve problems which cannot easily be solved by machines.

As human beings, we possess biases, preconceived notions, fear, ego, comfort zones, among other things; these can be detrimental to performing logical tasks. I have seen hiring managers and interviewers looking for experience in the exact field for which they are hiring, often overlooking candidates from related fields or with similar experiences. Let’s not forget the way some interviewers behave; if someone fuels their ego, they might even get hired. The world is so vast that the interviewer and the interviewee might never meet again, but arrogance can topple the conversation when someone has the upper hand and holds decisive power over another person. Additionally, there are many factors that can contribute to poor performance in an interview, such as health issues, a toxic environment, or harassment. However, such considerations are often overlooked in the process.

I will never forget the words, "Unsupervised NLP starts with LDA. If you didn't try LDA, what did you even do?" There is no one-size-fits-all solution in machine learning or data science, but this has become a pattern. In some interviews, candidates are questioned on some of their past projects. But all of a sudden, the interviewer becomes the expert on the subject and acts as if they are the authority on it. Despite the candidate having spent more time on and being more familiar with the problem, they may still be perceived as less knowledgeable due to the interviewer’s bias or preconceived notions. At times, the algorithm and its impacts are confidential, yet there seems to be no one willing to listen. Excuse me, but I will not disclose confidential information, even if doing so would secure the position for which I am interviewing.

Imagine a scenario in which a manager leaves a company and invites their subordinates to join them. The number of subordinates who would seriously consider the offer reflects the managerial qualities of that person. Similarly, if an employee is considered for rehire after leaving, it speaks volumes about their work and the value they contributed. I have seen people get rehired without a proper interview, yet they are rejected when explaining their previous work to other companies as a candidate.

I have witnessed individuals excel in interviews only to underperform in their roles, while others, who may not have interviewed as strongly but were given an opportunity based on intuition, have proven to be exceptional performers. Although interviews are flawed, some shortlisting criteria for candidates must exist. Recommendations are helpful, yet they can also be biased. The work history and CV provide an overview, but they can be easily falsified. One can cheat over some take-home assignments. I believe the best way to assess a candidate’s capability is to allow them to work on actual problems and observe the outcomes. Internships and contractual positions are possibly the most effective methods for hiring. Some companies are known for their hire-and-fire policies, but often overlooked is the development that occurs between hiring and firing. Candidates would likely be better prepared mentally if those positions were contractual. Contracts can be extended or converted to a full-time opportunity.

Lastly, I remain optimistic about the advancements in AI and hope that they may eventually make traditional interview processes obsolete.