I started looking for a new job passively in June 2026, having a feeling that some changes were coming. Turns out I was right: I got laid off in September.
I’d spent the previous few years working on blockchain data infrastructure and wanted to stay in roughly the same field. So I focused mostly on backend roles in Web3, particularly blockchain infrastructure, indexing, and data processing.
I found a new job in October, though outside the crypto industry. Over those few months, I went through quite a few interviews, received all kinds of feedback, and got a chance to compare this job search with the one I’d gone through three years earlier.
2023 vs. 2026
I was also laid off in 2023 and had to look for a new job. Both times, I was searching as an immigrant, which added another layer of stress. When your job is tied, at least to some extent, to the stability of your life in another country, the whole situation feels a bit different. Both times, I was looking in roughly the same niche: Web3 and blockchain infrastructure. The companies were different, of course, but the type of roles I was targeting hadn’t really changed.
Back in 2023, I was an experienced backend engineer but had relatively limited blockchain knowledge. I got some help with my CV, but didn’t put too much effort into it. Sometimes I’d show up to interviews without properly researching the company, or even without preparing questions for the interviewers. Looking back, those were pretty stupid mistakes. To be fair, I didn’t make them with the company that eventually gave me an offer. Still, I managed to land a job at a crypto company I genuinely wanted to work for, through a regular cold application. I ended up spending several great years there.
In 2026, I returned to the same market with solid experience in blockchain data infrastructure, indexing, DeFi integrations, and distributed systems. I prepared much more thoroughly for interviews, had a better understanding of how to present my experience, and actively used AI throughout the process.
You’d think finding a job would be easier this time. In practice, it was quite the opposite.
How I approached the job search
I used AI at pretty much every stage.
CV
I ran my resume through ChatGPT, asked it to roast it properly, and it helped shape it into something recruiters could easily understand and that would hopefully make it through automated screening.
Interview preparation
Before calls, I’d research the company, its product, the interviewers, and their professional backgrounds. I’d ask ChatGPT what questions would be worth asking and which parts of my experience I should emphasize. Sometimes I’d switch to voice mode and do mock interviews: ChatGPT would ask questions, I’d answer, and then we’d review my responses. Pretty useful, although it doesn’t help much with practicing live coding.
Finding vacancies
LinkedIn turned out to be a pretty useless job search tool for me, and so did most job aggregators. Too much noise, outdated listings, and questionable matches. So I took a different approach. I gave ChatGPT a detailed description of my professional background, tech stack, experience, salary expectations, and the kind of companies I wanted to work for. I also listed specific companies I was interested in and asked it to find similar ones. Then I set up a daily task: every morning, ChatGPT would check company career pages and send me relevant openings. I specifically told it not to trust aggregators: if a vacancy appeared on a third-party website, it had to be verified against the company’s official job board. Basically, I ended up with a personal recruiter who doesn’t try to sell me yet another Senior iOS Developer position just because I happened to write for iOS years ago.
Recruitment scams
Another thing worth mentioning is scams. In 2026, I was contacted quite often by fake recruiters offering suspiciously attractive positions. I barely encountered anything like that in 2023.
The patterns were pretty similar. Someone contacts you claiming to be a recruiter from a well-known company or agency. Or they represent some new startup whose description contains every trendy buzzword imaginable: AI, Web3, decentralized agents, and so on. Naturally, they have an amazing position, impressive compensation, and you’re the perfect fit.
What happens next varies. Some immediately offer a take-home assignment, while others arrange something resembling an interview first. Eventually, though, they send you a repository and ask you to clone it, install dependencies, and run the project. The repository might contain a perfectly plausible application, along with malicious code that executes during dependency installation or when certain scripts run. With Node.js, this often involves npm lifecycle scripts. You can end up with a backdoor on your machine instead of a completed take-home assignment. I almost fell for this once myself. It was my first encounter with this particular scheme, and I had no idea how common it was.
After a while, I started recognizing the patterns: similar opening messages, suspiciously attractive conditions, companies with questionable backgrounds, and a strong desire to get you to run someone else’s code as soon as possible. Of course, a take-home assignment involving an existing repository isn’t automatically a scam. I’ve had perfectly legitimate interviews where companies provided their own codebases. These days, though, I follow a simple rule: never run code from an unfamiliar repository without checking it first. If someone sends me a suspicious project, I download it as an archive without running anything and pass it to ChatGPT or Claude, asking for a security audit focused on malware, backdoors, and suspicious scripts.
LLMs are pretty good at this kind of initial analysis. Sometimes they can not only identify malicious code but also explain how it executes, what it does, and where it sends the data. It’s still worth checking package.json, install scripts, Git hooks, dependencies, and other possible execution points yourself. LLMs can miss things, so a clean result doesn’t mean the project is safe. If something still looks suspicious, don’t run it. If you genuinely need to complete the assignment, use an isolated environment without access to your secrets or main system. And if you find malware, report the repositories and recruiter profiles. People are just trying to find jobs. They shouldn’t have to deal with this crap.
What about a portfolio?
Turns out pet projects still matter.
Some colleagues tried to convince me that, with my level of experience, nobody would care about my portfolio anymore. My experience suggested otherwise. Mentioning your own project in an application or interview, especially something that actually works and people can use, rather than linking to a moldy five-year-old repository, can still earn you a few extra points for Gryffindor. Or whichever house you prefer. In several interviews, my projects became a separate topic of discussion.
Plenty of interviews. Not so many offers.
Getting responses to applications wasn’t a problem. Companies replied, invited me to interviews, and some even reached out to me first. I went through technical interviews, system design rounds, live coding, take-home assignments, and conversations with hiring managers and engineering leaders. I made it to the later stages quite often.
One major crypto company that had reached out to me and was pretty persistent about getting me into their interview process eventually rejected me with the feedback: “We’re looking for someone a bit more senior.” I asked for more detailed feedback and got a response from the hiring manager. The team hadn’t seen strong enough signals of ownership. At the same time, she explicitly said she didn’t doubt I had that experience — I just hadn’t demonstrated it well enough during the interviews.
At another company, I went through several rounds, completed a take-home assignment, and discussed my solution with the engineers. Everything seemed fine, but they eventually decided they needed someone with deeper Rust expertise. A third company wasn’t happy with how actively I used AI in development. Apparently, running two or three tasks in parallel with AI agents wasn’t ambitious enough. They wanted more. Some companies told me everything went well after the final round but decided to move forward with another candidate. Others simply stopped responding.
You can make it through a significant part of the process, receive positive feedback, and still find out that the company isn’t just looking for a strong engineer. They want someone who matches a very specific set of expectations almost perfectly.
My impression is that companies have become much pickier about candidates over the past three years. I have a theory about why. It might have something to do with the changing economics of the crypto industry. The days when money was flowing freely into Web3 seem to be behind us. A significant amount of investor attention has shifted toward AI, and crypto companies have to be more careful with spending, including hiring. That might explain the desire to find someone who checks every box and can start delivering value from day one.
But by searching for the perfect candidate, companies risk ending up without engineers who are perfectly capable of solving their problems. There’s a pretty big difference between “doesn’t meet every single expectation” and “can’t do the job.”
Live coding and AI
Companies had wildly different attitudes toward AI during interviews. Some strictly prohibited it. Others not only allowed it but effectively required it.
Traditional live coding is straightforward enough. With AI, though, you need to demonstrate how you break down a problem, plan the work, write instructions for agents, monitor their progress, and review the generated code. All within a limited amount of time. Some companies are already evaluating not just how well you write code, but how effectively you can organize the process of writing it with AI.
There doesn’t seem to be any industry-wide agreement yet on what that should look like.
What actually worked
Despite all the applications, interviews, take-home assignments, and multi-stage hiring processes, both offers I received came through people I’d worked with before. I accepted one of them.
Cold applications weren’t pointless. I met interesting companies and engineers, gained useful experience, and got a better understanding of the market. But if we’re talking strictly about getting an offer, good old networking won. In 2023, I’d managed to join a company I really wanted to work for through a cold application, with less relevant experience and much less preparation. This time, I ended up taking a job outside crypto, at a company working on applied AI.
I genuinely enjoyed working on blockchain infrastructure, but applied AI is an interesting field too, and I’m excited about the company I’ve joined. Given what’s happening in the crypto market, this might actually turn out to be a pretty good move.
What I learned
First, I don’t think I’ll completely stop keeping an eye on the job market again.
That absolutely doesn’t mean I’m constantly looking for a way to escape my current company. At my previous job, I got so absorbed in the project that I practically stopped paying attention to what was happening outside. Later I went through old emails and LinkedIn messages, saw some interesting opportunities I’d missed, and felt a little sad about it. Besides, if you wait until a layoff happens, you might find yourself competing for the same positions as a whole group of former colleagues who were on your team just yesterday.
This is especially relevant for people who, like me, live outside their country of citizenship. Even when your paperwork is in order, losing your job as an immigrant adds uncertainty to a situation that’s already stressful enough.
So keeping your CV up to date, following interesting companies, and occasionally talking to recruiters, even when you’re not planning to change jobs, seems like a reasonable strategy.
Second, you don’t necessarily have to wait for a company to post a relevant opening.
This time, I tried sending an open application to a fairly large crypto project I was interested in. They invited me to interview, and I made it pretty far into the process. Eventually, the company decided not to move forward for the time being, although they mentioned several times that I was still on their shortlist. Maybe they were just being polite. But other companies generally weren’t throwing that kind of language around. So proactively reaching out to companies you’re interested in is definitely worth trying.
And my main takeaway: job hunting is a skill of its own, and it’s worth keeping it sharp.
You can be a good engineer, have plenty of experience, and perform well in technical interviews, but none of that guarantees you’ll find a suitable offer quickly.
In 2023, I managed to break into Web3 with relatively limited blockchain knowledge. In 2026, I returned to the same market with several years of relevant experience, much better preparation, and a whole arsenal of AI tools.
And somehow, finding a job was still noticeably harder.