The lowest point wasn't the rejection from Cloudflare. It was two weeks after that, when I drove to my brother-in-law's place for Thanksgiving and lied to his face about how the job search was going.
"Oh yeah, things are moving. Just evaluating a few offers." I said it with the confidence of someone who had received zero offers and had a phone screen with a mid-sized SaaS company scheduled for the following Monday that I was quietly terrified about. That was November 2023. I'd been searching since June.
Let me back up.
Why I Was Even Looking
I'd spent three years at a fintech startup in Austin doing infrastructure work — Kubernetes on EKS, Terraform for basically everything, CI/CD pipelines in GitHub Actions and occasionally Jenkins when someone made questionable architectural decisions before I joined. The work was good. The team was small. I learned a lot.
Then the company went through a round of layoffs in May 2023, and I survived the first cut but not the second. Forty percent of the engineering org, gone in an afternoon. I had thirty days of runway on my employment before the severance kicked in, and I spent roughly twenty-nine of those days in denial.
By June I was actively job hunting. I figured: three years of real DevOps experience, comfortable with AWS, decent with GCP, could hold a conversation about observability without embarrassing myself. How hard could this be?
Very hard, as it turned out.
June–August: The Confidence Phase (And Its Collapse)
The first two months were almost encouraging. I got responses. I got phone screens. I got through a first round at HashiCorp (remote, which I wanted badly) and made it to a technical interview at a Series B company in Denver before stalling out.
The HashiCorp process was humbling in a specific way. I'd used Terraform for years. I thought that meant I'd be comfortable in their technical rounds. But their questions weren't just "do you know Terraform" — they were "do you understand the architectural decisions behind how Terraform works, and can you reason about them under pressure." I fumbled a question about state management in multi-team environments and I could feel the conversation shift. They were polite about the rejection. It still stung.
The Denver company rejection hurt differently because it came after a three-hour technical take-home that I'd spent an entire weekend on. No feedback. Just a form email that said they'd decided to move forward with other candidates.
August ended with four rejections and one ghost. I wasn't panicking yet, but I was starting to question whether my skills were actually marketable or whether I'd just gotten lucky with my previous job.
September–October: The Spiral
This was the bad stretch.
I applied to somewhere around thirty positions in these two months. I got twelve responses. Most of those didn't make it past the recruiter screen. The ones that did — a couple of companies I was genuinely excited about, including Datadog and a well-funded infrastructure startup out of New York — rejected me at the technical stage.
I started doing that thing where you refresh your email constantly and then feel slightly sick when there's actually something there.
The Datadog rejection was the Cloudflare I mentioned at the top, in terms of emotional weight. I'd prepped for it. I'd done mock system design interviews with a friend. I'd reviewed my distributed systems knowledge. But their technical bar was high, and I think I tried to sound smarter than I actually was in the system design round instead of just reasoning through the problem clearly. I overcomplicated a question about designing a metrics ingestion pipeline and then couldn't pull it back when the interviewer tried to redirect me.
I also bombed a behavioral round at a company I'd rather not name, which was embarrassing in a different way because I'd assumed that part would be easy. It wasn't. I rambled. I gave vague answers. I used the phrase "we achieved significant improvements" without any numbers, which is exactly the kind of answer that communicates nothing to anyone.
Around October I started actually changing my approach.
What Changed (Honestly)
A few things shifted in parallel, and I can't cleanly separate which one mattered most.
First, I started doing structured mock interviews instead of just reading about interview prep. I'd been using Pramp here and there for peer mocks, which was useful but inconsistent — you'd sometimes get paired with someone who was also struggling and the sessions didn't push me hard enough. I tried Interviewing.io for a paid mock with a Stripe engineer, which was expensive but worth it. That person pointed out that I had a habit of narrating what I was about to do rather than just doing it, which ate up time and made me sound less confident than I probably was.
I also spent some time with AceRound AI for async practice on behavioral and situational questions, specifically because I could do it on my schedule at weird hours when I was anxious and couldn't sleep. It's not a replacement for talking to a real engineer who's going to challenge your assumptions, but it helped me tighten answers and stop using filler phrases I didn't even realize I was using.
Second, I got more honest about what I was actually good at. My Terraform knowledge was solid, but my Kubernetes expertise was more "functional" than "deep." I'd been applying to roles that wanted a Kubernetes expert and then trying to fake my way through technical rounds, which wasn't working and wasn't really fair to anyone. I narrowed my focus to roles where the job description actually matched what I knew, which meant fewer applications but better fit.
Third — and this took me too long to admit — I asked for help. I had a former colleague who'd gone through a DevOps job search the year before and landed well. I bought him coffee and made him tell me everything. He pointed out that my resume had responsibilities written where it needed outcomes, and that some of my listed experience was vague in ways that would make a hiring manager nervous. I spent a weekend rewriting it.
November–December: Things Started Clicking
The Thanksgiving lie I told my brother-in-law came right before a stretch where things genuinely started moving.
That Monday phone screen I was terrified about went well. The company was a mid-sized logistics tech firm — not the sexiest brand name in the world, but the infrastructure problems were interesting and the team seemed thoughtful. I made it through three rounds in about three weeks, which felt fast compared to everything that had come before.
Simultaneously I had a process going with a fully remote infrastructure-as-a-service company that had found me through a referral. That one moved slower. By mid-December I had two offers on the table within the same week.
One paid more. One had better scope and a team I'd enjoyed talking to. I took the one with better scope and a slightly lower base, which felt like a very mature decision and also made me briefly question whether I was being an idiot.
I started in January 2024.
What I Actually Learned
Six months felt like forever while it was happening. Looking at it now, I think it was mostly a calibration problem on my end — I had real skills, but I was applying in ways that didn't reflect them accurately, and I was doing interview prep that felt like work but wasn't actually making me better at interviews.
The behavioral stuff being a weak point surprised me. I'd mentally categorized that as the easy part and focused almost entirely on technical prep. But a lot of DevOps interviews care deeply about how you handle incidents, how you communicate with non-technical stakeholders, how you've navigated disagreements about architectural decisions. Those questions require actual stories, not abstract explanations of what you'd hypothetically do. I didn't have my stories organized, and it showed.
The companies where I made it furthest were the ones where I'd done the most specific research — not just "here's what they build" but "here's a problem they've probably encountered at their scale, and here's how I'd think about it." That prep is tedious and doesn't transfer between companies, which is why most people don't do it. I think it's also why most people plateau at the second or third round.
Would I have gotten to an offer faster if I'd started with a more structured approach in June instead of coasting on confidence? Probably, yeah. I lost maybe two months just assuming I'd interview my way through on experience alone.
The job's been good, for what it's worth. The infrastructure problems are real, the team argues about the right things, and I haven't lied to a family member about my employment situation since Thanksgiving 2023. That's enough.
Want real-time interview assistance? AceRound AI works live during Zoom/Meet interviews.
Top comments (0)