Engineering video
8:25 · 23 July 2026
Senior developers using AI tools find that they write a high percentage of code with them, but these tools require deep fundamentals, constant guidance, and extensive rework, proving that they are meant for experts rather than replacing developers.
Watch section from 0:26 →The idea of an autonomous agent working overnight is practically unworkable because models overoptimize or randomly improvise when asked to improve architecture, requiring human judgment and intervention.
Watch section from 1:43 →Over the past five years, salaries have largely stagnated despite real inflation, leading to a more crowded market with fewer jobs where candidates compete for the same pay against rising expenses.
Watch section from 2:56 →Finding a job now requires a structured, data-driven approach where candidates track their application numbers like a scientist, treat the search as a full-time job, and avoid emotional attachment to single companies.
Watch section from 3:39 →Candidates must take charge of interviews by establishing a clear roadmap, managing time across phases like requirement gathering and minimum viable systems, and providing concrete evidence of seniority.
Watch section from 4:49 →Hey guys, here from the scene with together with and here in Amsterdam. So yesterday we had the and get together with and of course there was a lot of back and forward and lot of questions. And the main question was of course about hey, what's going on in the job market? What's happening with AI? What's the future of software engineering and JavaScript? We've seen LLMs and AI popping up and reading more and more of the developer job.
There's one point for sure companies are saying, hey, we're going to use AI and we don't need any means in junior, what's your take on it? So what we've seen with the with is the ones that are employed, they are not afraid of AI. And those people they work with it, they are in the trenches, they work with cloud code, most of them. A lot of them are using fable, which is the latest model. And what I notice is that the more senior they are, the least worried they were about AI.
Because they kind of push the model to the max, a lot of the mark token max thing, most of them said that they are writing around 80 to 100% of their code with AI, and still they were not worried. And the reason was because they realize how much rework you need to do, how much you need to guide the model and they realize that you need senior knowledge. And so the way they see those tools is that they are just tools for experts. So you need to be an expert.
And one of our mentees, she just landed a senior front end job and she told me, but I felt that the fundamentals are now more important than ever. That were her little words, she said, you cannot take any decision if you're not really deep into software engineer. So they were not worried. But the people that are actually in the market, the ones that are interviewing, they are afraid because they get rejected, or because they get a certain messaging, because the salaries, you know, they kind of stagnated and companies are using the AI narrative.
So depending on where you are, you might feel differently about AI. But one thing I want you to expand on and it's something that we've seen other than the more we work with AI is that those tools are expert tools, like you mentioned. What do you mean by that? And can you expand on that on that concept? So the number one thing is that everybody agreed that whenever you give it a prompt and you just want a new UI, the model is very good.
But the moment you want to tweak things, or you want to clean up the code, that's when you need judgment, and if you tell the model, hey, can you just write clean code or can you improve the architecture? It will randomly. improvise or overoptimize towards that prompt. So a lot of them told me, it just, you know, you got to fight with it a lot. One of the Menti's, she actually built her own and told it, you know, exactly how it needs to code and then it still needs rework.
And they told me that this dream of the autonomous agent that works overnight, that's always, almost never a workable result. You always have to go in, you got to tweak it, and then you go back to fundamentals because you look at the way it writes CSS or JavaScript. And if you don't know what you're doing, you know, you're really at the mercy of whatever the model gives you. So to sum it up so far, number one, AI it's an expert tool and we've seen the more senior you are, the less worried you are, fundamentals matter more than ever.
And we are the senior dev, we had a huge like, you know, to this last six years, we're always focused on the fundamentals, fundamentals. When this AI thing kind of popped up and we had to move towards that. We need to have an opinion on it, but our were telling us, hey guys, the reason why we like the and the reason why we came, it's the fundamentals itself. So the second point would be master the fundamentals and finally, what about the job market?
Yeah, so that's where things got a bit more complicated in this last five years, where a lot of people complain that salary is kind of stagnated. So if you look at, I think we discussed with one of them in the local inflation, he announced them, it was around 20%, the real inflation. But when you look at the salaries over those five years, they're pretty much at the same number, like you usually get, you know, the mid level.
He announced usually around 80 90k and then the seniors kind of 110, but that's what it used to be. So we were talking about pre-changed power with a lot kind of saying that, you know, you you have to compete in a market where it's a bit more crowded. Where there's a bit less jobs and you compete for the same salary while your expenses went up. Now I'm not going to get into detail about why that happens, but what I've seen with most of them is that, you know, everybody agrees that five years ago they used to get a job with kind of an average resume.
You interview with a bunch of companies and you get the job, nowadays you got to really structure and systematize your job search. And I think the doc agrees with us that you have to be very systematic about how you go and apply to jobs. between top tier companies, even he knows that where they pay way over 120k and then the average salary, so if you want to make the jump, you really need to be very serious.
And what we all agreed is that looking for a job, it's a full-time job. Bogdan. So, I'm a software engineer, I'm in the market, you know, I need to find a position. What would be the the main four or five key points that you will tell me? So what we learned talking to one of the software engineers that we mentor here, uh, she's been in the market, she got an offer, and there's basically three things that really made the difference.
And so number one was, you really want to systematize your job search. You don't want to be emotional, you need to track your numbers. It's a very intense process, it's very emotionally draining, so you need to look at the numbers. You need to be a scientist because you're going to have very bad weeks and you're going to have very good weeks. So, number one, you need to track your numbers, you need to track your applications, you need to track your interviews.
Number two, in the technical interview. You really need to lead the interviews, and you need a clear structure, but at the same time, you need to make sure that by the time the interview finished, there is enough evidence for you to pass that interview. And she literally told me in this this uh one of the three offers she got, she was able to close that interview because she was always looking at the clock and she kind of leaded the uh the the interviewer and said, look, you know, um we're here, we're in the requirements getting requirements getting phase.
We are are going to go all the way to final design, did we lead enough? Can we move on? So you really need to lead the interview, especially in system design interviews. And number three, do not fall in love with any specific company. So what happens is when you are in the market, you get rejected and, you know, let's say you got rejected three times and what it's happening in your cost and then there's this one company that treats you well, you're kind of emotionally vulnerable, you want the positive answer, and then people project and they want this one company, they forget about their job process, they put all their eggs in one basket.
You might get rejected even at the office stage, and then what? And then they are wrecked, they are destroyed, and then, you know, it takes them three months to get back to it. So, again, you got to be very systematic and keep your emotions in check. really big numbers driven, data driven there. One more point, now that we talk about the design market, we are in a senior only market. We know that companies are mainly looking for senior engineers because they can use AI. And you mentioned something about the doing the tingle interview process, you have to make sure that you're putting together enough evidence, enough proof to signal that seniority. Like my question for you, it's how can I what's the best way like for me to signal seniority to all the different tingle interview stages. So to give you a concrete example, this was a front system design interview. And what she did really well and what we teach the senior devs, she really let the interviewers know what is her roadmap for the interview. So you come in and even if they're interviewing you, you set the stage because they're looking at you, they expect you to kind of make it happen. It's like you hire a chef, they need to know how to use everything, you know want to tell them how to do this, do that. So she did that very well, she structured it really well, and then she was able to push them forward. Because what happens in a system design interview, and we see this a lot on, they're like, oh, we are in the requirement gathering phase and they spend three hours there, the interviews over and you're done. Got it. So you really want to be able to get people through all the stages so from requirements gathering to the minimum viable system and scalability. And leave, you know, leave drawings, if if it's a whiteboard interview, make sure you draw, they need to get something with them from the interview. Exactly. You know, writing things down, if it's a call, uh this uh, you know, if it's in person, it's it's usually a little bit easier to interact with them. But you need to create something that she said it really well. She said at the end of the interview you want to have a working prototype. You know, if it's a live coding interview, the code has to run. If it's a system, the system you have to be complete. And and it makes sense if you think about it, like the market is very competitive. The companies need to know that if they hire you, you're going to be able to come into the team and deliver value from day one. It's all about value. It's not the 2021 market where it was enough to just have some some buzzwords and title. Um, thank you very much, Bogdan, for us, the next top, it's going to be London where we're going to see a lot of more of our. And by the way, if you want to find out where you are in the seniority scale, they say free.
pink application in the comments. Click there, it's built with real interview questions that we gathered over the years. And it's going to tell you if there's a specific gap in your knowledge, so check it out in the links in comments. And finally, if you want Bogdan Knight to help you with a personalized code map, live feedback and an amazing community, and maybe we get to see each other in the city you live in, then check the link below, you'll see how you can work with us.
Thank you, Bogdan, and we will see you folks in the next one.
