Take every interview they'll let you take

At a big company, taking interviews is treated as a tax. It lands on the senior engineers, it eats an afternoon, and the reward for doing it well is that you are asked to do more of it. I came to see it the other way round: volunteering to interview was the smartest, least glamorous investment I made at PhonePe.

None of what it paid back is on your team's roadmap, which is exactly why it is easy to skip. Here is what it actually bought.

The rounds I took

Over my time there I took part in interviews across all three formats:

• 50+ DSA problem-solving rounds: watching how a person thinks when they are stuck, which is the part that actually tells you anything.
• 10+ machine-coding rounds: watching them build something small and working against the clock.
• 10+ design rounds, shadowed: I was cleared to run these on my own, but chose to sit in on the senior engineers first, saying little, watching how they worked.

The two coding rounds kept me sharp more than they taught me. A DSA round runs a standard arc (set the question, agree the approach, code it, state the complexity, run the tests), and most candidates walk it the expected way; the occasional non-standard approach was the one worth being there for. Machine coding was much the same: now and then a candidate solved it a way I had not considered, and widened how I saw the problem. Neither taught me a great deal: both kept me in touch with the fundamentals.

Most of what I actually learned came from the other two, the design rounds and the campus drives: high-pressure, deliver-now environments.

I started with very little exposure to system design, so there was a great deal to learn, and shadowing seniors is where I learned it. Watching someone senior probe a design shows you what a good one sounds like:

• What they push on, and what they let slide.
• How they follow up, turning each answer into the next question.
• When they stop, the moment they already have their signal.

You learn the round from the far side of the table, and from there you see what you never could as a candidate:

• The mistakes that sink people over and over, and how avoidable most of them are.
• The pitfalls a candidate walks straight into, and the interviewer watches coming.
• How the strong ones drive it: scoping the problem before they touch it, thinking out loud, steering toward the ground they know best.

Sit through enough of them and a pattern forms that you could never assemble from your own handful of interviews. There is nowhere else to learn it.

The nearest thing money can buy is a mock interview, and people pay real money for a handful of them. I had the equivalent for free and without limit: every round I sat in on, and later every one I ran, was another rep at the exact skill. That is a luxury, and I knew it.

The golden rule

Campus hiring at PhonePe was accelerated learning. The bar was high. To keep up, you had to revise and recall the fundamentals fast. In a single day you would work through fifteen to twenty problems, most above average and a good number on the tougher side, material I would have taken weeks to get through at my own pace. The night before a drive I would be up until two or three in the morning, not revising trivia but formulating every approach a candidate might reasonably bring to a problem, and working out where each one broke. And not just me: everyone taking part in the drive was up late, doing the same.

That preparation exists to serve one rule, the rule I ran every round by: if you cannot refute a candidate's approach, you have to accept it. You do not get to reject an answer because it is not the one you would have written. You reject it only if you can point to a concrete flaw, and if you cannot, it is valid, and you say so. It keeps you honest, and it is unforgiving of your own preparation: to judge an approach fairly on the spot, you have to have already thought it through yourself, the night before, at 2am.

It is also what makes the round land for the person on the other side. A candidate once told me, at the end of a round, that it was the best interview they had given in their life, and you do not get that by grilling anyone. You get it by having done the work, so that the round is hard and fair at once, and the candidate can feel both.

The campus drives

The drives took me across the country, with a few (like Cummins College) run online.

I will admit I was strategic about which drives I put my hand up for. A campus visit is a fully sponsored trip, so my first filter was simple: does the city have an airport, so the travel is easy? After that, I optimised shamelessly. Where I went, and why:

PEC Chandigarh is home: the drive doubled as a fully sponsored trip back to my own city.

NIT Surathkal is the only campus in India with its own beach; coming from BITS Goa, I wanted to settle the beach debate with my own eyes.

IIIT Hyderabad is regarded as one of the best in the country for computer science, a no-brainer.

IIT Kharagpur is the oldest of the IITs, and still among the best for CS.

SPIT Mumbai was a chance to see a friend in the city.

MNNIT Allahabad is the alma mater of my friend Sujeet Kushwaha.

By the end, the recruiter running the drives knew me by name. I had turned up that many times.

On one of the hiring drives, I was sharing a room with Ketan Daithankar, and we stayed in touch long after the interviews were done. We never worked together (different teams, always), but across my whole time at PhonePe his guidance was instrumental, a lifesaver every time. I'm forever grateful for all of it.

What it paid back

Three things, and none of them is the reason you get asked to interview in the first place.

• A network that unblocks you. When you have met someone in person, they reply faster. A cross-team dependency that would otherwise sit in a queue for days moves in an hour, because you know the person on the other end and they know you.
• The gossip nobody puts online. Interviewing across teams, and meeting the people on them, got me the unwritten version of the company: the candid, word-of-mouth read on what each team is really like to work on, the day-to-day that never makes it onto a review site or an org chart. It is exactly the intelligence you wish you had before you choose where to sit.
• Exposure to the whole company. This is the one I value most. You come away understanding the place far past your own service. Most engineers spend a career inside one team and never once see how the rest of it is built; interviewing drags you across all of it.

What I'd tell you

• Say yes to interviewing. The craft, the network and the exposure compound, and none of it competes with your real work for roadmap space; it simply is not on the roadmap at all.
• Shadow before you solo. Being cleared to run a round is not the same as knowing how to run one well; the fastest way to learn is to sit beside someone who already does, before you ever run it alone.
• Out-prepare the room. The golden rule (refute it or accept it) only holds if you have already thought through every approach before the candidate sits down. The 2am preparation is not overkill; it is the job.

None of this was the job. All of it made me better at the work, and better at knowing what to look for next time.