Download AppSix minutes left on the clock, and I still hadn't figured out why my brute-force solution kept timing out in my head before I'd even written a line of code. The interviewer hadn't said a word in almost ninety seconds, just watching the shared screen, and I remember thinking, very clearly, that this was the exact moment where every other application that season had already quietly failed me and I couldn't let this one go the same way. That moment came about three weeks into a process that, looking back, taught me more about how a quick-commerce company actually thinks than any amount of reading about Zepto beforehand ever could have.
Let me back up to where this actually started.
How a Google Form on LinkedIn beat every formal application I sent
I'm from a tier-2 engineering college outside the usual campus-placement circuit for a company like Zepto, so this was always going to be off-campus, no referral, no warm intro. I'd applied to a stack of companies that season through the usual portals and gotten mostly silence back. The Zepto application, oddly, came through a Google Form a Zepto employee had posted on LinkedIn, the kind of link you bookmark at 1am half-convinced it goes nowhere. It went somewhere. A recruiter's email showed up eleven days later asking if I was still interested, by which point I'd genuinely half-forgotten I'd filled it out.
I'll admit a chunk of my applications that season were MAANG-shaped, the safe, obvious targets everyone in my batch was chasing. Zepto wasn't originally on my radar until a senior from my college, two years ahead of me, mentioned the pace of execution there was unlike anything he'd experienced at his previous service-based job. Quick-commerce sounded chaotic in a way that scared me a little, and also, if I'm honest, excited me more than another slow-moving enterprise SaaS pipeline would have. Before applying I spent a couple of evenings reading around, less about interview questions at that stage and more about the company itself, what it had built since starting out as KiranaKart, how the dark-store model worked logistically. It wasn't strictly necessary prep, but it meant I wasn't walking in with only a vague, borrowed sense of "fast-paced startup."
That senior's advice stuck with me in a way I didn't fully appreciate until much later. He'd said something like, "don't expect them to slow down for you just because you're a fresher," and at the time I nodded along without really internalizing it, the way you nod along to any piece of secondhand career advice. It took actually sitting inside a twenty-five-minute coding round to understand he hadn't been exaggerating for effect, he'd been describing the literal, minute-by-minute texture of the process.
The recruiter call that quietly set the tone for everything after
The HR call was short, maybe fifteen minutes, mostly logistics, current expectations, notice period questions that didn't really apply since I was a fresher. Two technical rounds on a platform called BarRaiser, then a hiring manager round if I cleared both. No online assessment for my batch, which surprised me, I'd been bracing for a HackerRank-style test the way most product companies do it. She mentioned, almost in passing, that the technical rounds move fast and I should expect roughly twenty-five minutes per coding question, not per round, per question. I didn't fully register how literally she meant that until I was sitting inside round one myself.
I used the ten days before round one almost entirely on speed rather than new topics, drilling patterns I already knew until I could get moving without the usual few minutes of just staring at a problem statement. I also read through whatever write-ups I could find online that week, mostly LeetCode Discuss threads and a couple of Medium posts from people who'd been through Zepto's process before me, and one detail kept surfacing across completely unrelated posts, independent strangers describing the exact same twenty-five-minute pressure the recruiter had warned me about. Seeing that pattern repeat did more for my mental prep than the actual grinding did, it meant she hadn't just been trying to scare me into over-preparing.
There was also a smaller, quieter kind of prep happening alongside the DSA drilling that I didn't recognize as prep at the time: I started timing myself on problems I'd already solved before, just to get a feel for what twenty-five minutes actually felt like in the body, not just on paper. Turns out there's a real difference between "I know I can solve this in twenty-five minutes" and "I have physically sat through twenty-five minutes with a clock visible and a stranger watching," and no amount of untimed practice closes that gap.
Back to that clock, and what actually happened next
So, round one. The interviewer joined on time, no extended small talk, straight into it, matching everything I'd read about Zepto's pace. First question was a sliding window problem dressed up in delivery-logistics language, tracking the maximum number of concurrent orders a rider could be assigned within a rolling time window. I recognized the pattern almost immediately and got through it with a few minutes to spare.
The second question is where that six-minutes-left moment from the very start of this post actually happened. A stack-based problem reframed as figuring out how long until a delivery slot's demand exceeds the current one, and the delivery-logic framing genuinely took me longer to untangle than the algorithm underneath it did. Once I finally saw the monotonic stack pattern hiding inside the wording, the coding itself moved fast, but I finished with maybe ninety seconds to spare and slightly shaky hands typing the last few lines. He said "okay, that works" and nothing more, which left me completely unsure walking out of that call whether the round had actually gone well or just gone.
The lesson I only understood afterward: the twenty-five-minute constraint isn't really testing your algorithm knowledge, plenty of freshers know these patterns, it's testing how fast you can translate an unfamiliar wrapper back into a pattern you already trust. Practicing solved-in-my-head problems under an honest clock beforehand would have shaved a good three or four minutes off that scramble.
The design round I almost walked into unprepared
Round two invite came two days later, which felt like an eternity given how fast everything else had moved. I'd assumed, incorrectly, that a fresher SDE-1 role wouldn't go deep into low-level design, and I was wrong. The interviewer asked me to design a simplified inventory and order-assignment system, the kind of thing sitting behind a ten-minute delivery promise, entities like Order, DeliveryPartner, Inventory, how they'd relate, a rough class structure, a couple of key methods.
I fumbled the first few minutes trying to jump straight to code instead of talking through entities first, and he gently redirected me: "let's step back, what are the core objects here before we write anything." That single nudge probably saved the round. Once I slowed down and actually reasoned about the domain, ownership, availability checks, what happens when two orders compete for the same inventory item, things clicked. The last fifteen minutes were a single sharp follow-up, how the design would change if inventory lived across multiple dark stores simultaneously. I didn't have a fully polished answer, but reasoning through it out loud seemed to matter more to him than landing on a memorized "correct" pattern.
I closed my laptop after that call and sat there feeling more drained than after round one, despite it technically being the "easier" round on paper. That gap between how a round feels on paper and how it actually feels in the room is, I think, the single most useful thing to know going in: don't assume seniority of topic maps cleanly onto difficulty of experience. A few hours with basic object-modeling practice beforehand would have meant I skipped that fumbled first minute entirely.
I also spent that evening messaging the same senior who'd first pointed me toward Zepto, asking if the LLD-heavy structure matched what he remembered from his own process. He confirmed it had been similar for him too, a design round that leaned more toward reasoning about a live operational domain than toward memorized design patterns, which made me trust the round a bit more instead of assuming I'd simply drawn an unusually design-heavy interviewer by chance.
Five days, a conversation instead of an interrogation, and a question that actually mattered
Five days of silence after round two, shorter than waits I'd read about for bigger companies but still long enough that I refreshed my email every time I finished a class. The HM round, when it came, felt like a different interview entirely, calmer, more conversational. He'd clearly actually read my resume beforehand, his follow-ups were specific to a side project, a small inventory-tracking tool I'd built for my college canteen, which suddenly felt oddly relevant given what I'd just been asked to design an hour of interviews earlier.
Near the end he asked directly whether I understood Zepto moves faster than most fresher-friendly companies, and whether that excited me or worried me. I said both, honestly, which got an actual laugh instead of the polite silence I'd braced for. He also asked what I knew about the company beyond the ten-minute delivery promise, and I mentioned the shift from Mumbai to Bengaluru as headquarters, the scale of the dark-store network, the sense that the engineering org was still small enough that individual decisions actually mattered. Answering honestly rather than performing the "I thrive under pressure" script, and actually having something real to say when asked what I knew about the company, mattered more in that round than any technical prep could have.
We also spent a good five minutes just talking about the team itself, what a typical sprint looked like, what shipped in the last month, which I only recognized afterward as him quietly checking whether I actually wanted this specific role or was just chasing the brand name the way plenty of candidates that season clearly were. I asked more questions back than I usually would in an interview, partly genuine curiosity and partly a half-remembered piece of advice about questions signaling ownership, and either way the round ran noticeably past its scheduled slot, which felt like a good sign even before the recruiter's call confirmed it.
The offer call, and the six minutes that finally made sense
Four days later, the recruiter called, straightforward and warm, congratulations, here's the offer. Without getting into exact figures since these vary a lot by year and negotiation, it landed in the range other fresher SDE-1 offers at Zepto that I'd seen discussed online seemed to cluster around, fixed base plus a stock component vesting over a few years plus a joining bonus, nothing shocking compared to what other well-funded quick-commerce startups were reportedly offering freshers that season. I said yes on the call itself, no need to think it over, which apparently surprised her slightly, she mentioned most candidates ask for at least a day.
I didn't need one, because that six-minutes-left moment from round one had already told me something the rest of the process only confirmed: the speed I'd been tested on wasn't a performance designed to stress candidates out, it was genuinely representative of how decisions get made there day to day. There's very little gap between who they hire and what the job actually is, which in hindsight made the jump from interview process to first week feel far less jarring than I'd braced myself for during that whole shaky-handed final minute back in round one.
There was still a gap of a few weeks between accepting and actually joining, background verification, documentation, the usual onboarding admin every company runs a fresher through, oddly ordinary after a process that had spent so much of itself racing a stopwatch. I used the gap to rest properly for the first time in months, which I hadn't managed once during the entire application season.
Looking back at that six-minutes-left moment now, from the other side of an offer and a joining date, it reads less like a crisis and more like the most honest preview of the job I could possibly have gotten. Nobody sat me down afterward and explained the culture in a slide deck, the interview process simply was the culture, compressed into three rounds and a handful of days. If you're weighing a company like this against a slower, more traditional one, that compression is worth paying attention to, not as a red flag, but as data: whatever the interviews feel like is roughly what the job will feel like, and it's worth being honest with yourself, the way I eventually was in that HM round, about whether that pace is something you want or something you're just telling yourself you want.
If you're prepping for something like this, the mock interview sessions on devsunite.com/resources are what got me comfortable actually talking through my approach under a timer instead of freezing the way I used to in earlier interviews that never went anywhere.