Thread by @ngkabra

He was almost depressed when he first came to me. He used to get shortlisted for campus interviews but then fail in the technical interview rounds. The year ended without him getting a job, and he became one of those zombies going from walk-in to walk-in with little luck. /1

Campus interviews are brutal and also broken. Companies reject students for all sorts of preposterous reasons.

A couple of rejections and the student goes into a downward spiral—rejections result in low confidence, and low confidence results in more rejections. /2

I could see that he was well into the downward spiral.

Before looking for a solution, I needed to know what the real problem was.

"In the interviews, what kind of questions are you doing badly on?" I asked. /3

"[Well-known company X] asked me the difference between B-Tree and B+-Tree. I didn't know that. [Well-known company Y] asked me to write code for some array processing that required recursion. I couldn't do it."

OK, so those were the weaknesses. What about the strengths? /4

Everyone has strengths. The tricky part is figuring out how (and if) they align with what industry is looking for, and how to highlight those.

I needed to find out whether he was, in fact, employable by the software industry, or did his strengths lie elsewhere. /5

I often run into students who manage to get into a BE(CS) course, and complete it with reasonable performance, but have no interest in programming.

For them, doing an MBA makes more sense than a programming job. Or in some cases, getting a job as a Business Analyst. /6

What category was he in?

"In Software/CS, what do you like? Have you done any substantial project?" I asked. /7

It turns out that in the break after graduation, he had taught himself machine learning and done some sentiment analysis of tweets from the American Presidential elections (the previous one).

And he had also built websites for the college tech events.

Jackpot. /8

Having this kind of experience/capabilities easily puts him in the 95% percentile of what companies need. Unfortunately, this is invisible to the standard filters applied in campus-hiring process because it is not common. /9

The hiring processes of most companies are broken due to a combination of one-size-fits-all filtering criteria, HR personnel who don't have the background to appreciate non-standard skills, and companies' lack of trust in their interviewers' judgement. /10

How do you win this game? You could try to play by the rules: study to get good marks, study to ace the aptitude test, study to ace the written round, and memorize standard answers to usual technical interview questions to ace the interview round /11

(Most students do this, and this is sad because it leaves very little time to learn the stuff that is actually needed on the job: writing and modifying non-trivial programs. But that's a story for another day.) /12

Clearly, he wasn't going to win this game playing by the rules, since he was bad at memorising answers for an interview.

He needed a different angle. /13

I asked him if he would put his sentiment analysis code on github, and build a website that shows the code running live, but also has a report with appropriate graphs, and explanation

"Yes," he said confidently. /14

2 weeks later, he was back with a working website.

I asked him if he could modify the website so that it was India focused, the user could enter any date range and the data would get updated on-the-fly, and the tweets could be filtered by the minimum number of likes. /15

That took another 2 weeks.

I sent a link to that site along with a 2-3 line introduction to a friend of mine in a company that was hiring.

An interview was scheduled 2 days later. /16

Imagine the situation so far.

Now answer this: In this interview, what do you think was his answer to the "B-Tree vs B+-Tree" kind of questions? And to the "write a program with recursion in it" kind of questions?

Try to answer the question before reading on. /17

Tricked you.

There were no B-Tree vs B+-Tree question. There were no questions involving recursion. Most of the questions were about how he built this website, the data-structures he used, and how he would modify the program to do X or Y. /18

I have a PhD in Databases, and I've been part of a team that built an actual database, and still, I've never in my career been in a situation where knowing the difference between a B-Tree and B+-Tree made a difference (except in my exams). /19

Companies ask these questions in interviews because they don't know what to ask. How are they supposed to choose between 100 students, most of whom have never really written a program more complex than Fahrenheit to Celsius conversion? /20

This is the street-light effect: They ask (irrelevant) theory questions because there are no relevant practical questions to ask.

But show them a substantial practical project, and the good companies will grab on to it like a drowning person grabbing a plank /21

If you can't win the hiring-process game, find ways to bypass it.

If you don't have the paper credentials to get past HR filters, or the theory knowledge to clear the default interviews, build a practical demonstration of your real, actually useful capabilities. /22

Epilogue: He got the job. And, a just few months later, an increment and a promotion. And 4 years later he's still there, doing very well.

(Note: this is a fictional composite based on multiple real-life stories I've been personally involved in.) /23

PS1: An individual side project carries more weight than a BE Project.

Companies assume that BE projects are done by just one team member and they're not sure you are that one.

Also, having a side project shows more passion than a project done as a necessary requirement. /24