2025 in Review - Perspective of a Junior Software Engineer
At the start of the year, the excitement I’d felt about my recent promotion quickly faded due to a radical series of events that threatened the job security of pretty much everyone I knew. I was a federal contractor then, so of course, we were at the front of the line when the layoffs hit. A clear message was sent to all employees and contractors alike: Nobody Was Safe.
Performance didn’t matter. Ongoing projects didn’t matter. Your reputation around the office didn’t matter. Eventually a pink slip, jargon-littered email, or someone from HR would call to inform you to leave everything at the office/campus and return home immediately. Much of the senior staff began to take early retirement, some forcefully asked to resign, and some not even that lucky. In the time since all this happened, much of the workforce has been reinstated, but understandably, shaken to their core.
Ebb and Flow
Strangely, I feel glad that things happened the way they did. While the promotion to Full Stack Developer felt great on paper, I was still holding onto all the responsibilities of my previous helpdesk position because we were understaffed. Those often took priority, so weekends became study-sessions for upcoming project requirements, or crash-courses on this weeks new stack(s) suggestions. And because I was the only developer in our team, I had no one to bounce problems off of. While leadership obviously cared and tried their best to work within the environment provided to us, bureaucracy became the cause of many frequent headaches.
There was no central repository. What teams existed within the organization were entirely compartmentalized and isolated while working frantically across different stacks with no standardization. When I needed help, I’d hunt down the nearest senior software engineer I could find, send a handful of emails, ask around, eventually get someone on a call… And the conversation always ended the same way: “Well, that’s certainly a lot that you’re doing. Could you tell such-and-such about this? Keep up the great work!” or “Hmm… have you been in touch with so-and-so? They know a lot about this”, before I’d be handed off to another person, another team, with their own set of problems and vague understanding of what this new AI thing available on Microsoft Azure™ was.
You Never Feel The Exponential
I, along with many others from open-source communities, have always felt innovations in deep learning to be majorly disruptive; For years, we’ve discussed the implications for the economy, many professions, education… you name it. One thing that I may have taken for granted, however, was just how disruptive it can be for people completely outside of these insulated inner-circles.
It’s one thing to spend your time around people who are developing toolkits utilizing all of the latest innovations in AI research. Chatting while others work with various frameworks, creating new architectures that run on consumer hardware to help them explore different problem sets. When everyone around you is constantly talking about these things, the speed of progress is felt in a very different, almost deceptive way.
Stepping outside of this bubble is a dissonant experience. The perceptions of the average consumer vary wildly across how they use it and their overall opinions of it. In most cases, it seems a large part of the public’s understanding is shaped by headlines and hype/FOMO cycles, over hands-on experience and rigorous testing. But, I think that’s okay?
While the bleeding-edge research is incredible, the consumer-friendly, go-to apps and interfaces are a shallow pond. How can anyone be expected to take it seriously when all they’ve been exposed to is the same chat interface and spam-bots? How are they to know what that even means, what a model even is, or where to source that information? Sure, they could ask the thing itself, but 200 samples is enough to poison an LLM, so who’s to say that you should take it at its word?
The consumer market is, in many ways, much like the federal contracting environment; it lags behind. It takes time for new research to be integrated into the norm. 5 years, 15 years, who knows. One thing’s for certain; while the roadmap needs a lot of heavy work, the economic impacts are already being felt by many people. How that resolves in the short/mid term, no one can be certain, and I feel immense empathy for those affected by all of the layoffs in 2025.
Learning to Ride the Spiral
Given the state of global acceleration, and life feeling like it was doing much the same, I stumbled into the last quarter of the year with a weird mix of burnout and hyper-vigilance. I clawed away at over 1,200 job applications in the span of 10 months, lost my apartment, and made the 9-hour drive back home.
It kicked the hell out of me, but it forced me to rebuild my standards and my sense of direction. I had to challenge everything I thought I knew about how to learn, what the future may hold, and how these innovations roll out in practice. And while short-term plans had to change suddenly, I still felt it was important to keep building something. I budgeted, purchased textbooks, online courses, forced myself to use Arch as a daily driver to become fully submersed in the open-source ecosystem, and configured my NAS and the few servers I have to support developing different tools/apps in my homelab.
Just as I was starting to think I’d burned through every job option, a good friend of mine reached out with a lead. This time it wasn’t a faceless HR clerk or some ghost-listing to collect your data; it was someone who has known me personally for years and thought I’d be a good fit. I accepted, albeit while feeling incredibly anxious. Having an opportunity to work with a team of other engineers on something that we can all be passionate about is a dream-gig. What if I choked during the interviews? Would I even work well with the rest of the team? “Backend Platform Engineer” sounds like a LOT!
Of course, this was the early onset of every SWE’s favorite feature: Imposter Syndrome. Had everything I’d been making over the years actually given me enough experience to solve these problems? Is it normal to question yourself regularly, or feel like you’ve completely forgotten about a problem space if you step away for a few days? Should I be asking this many questions? How many Google searches and files do I dig through before I ask the lead engineer about something? Are my questions focused enough or too broad?
Short answer: It’s all so fking normal. At this stage, my lack of experience will shine through. Being nervous about a code review is expected, feeling overwhelmed while you’re on-boarding is expected, but no one should be upset at you for asking questions that show you’ve at least put some thought into a problem. Turns out, there are SWEs with 10+ years under their belt who still feel like they’re just hanging on for the ride. The difference is simply that they’ve learned to embrace challenges as they appear and excel at breaking them down into solvable pieces.
That’s how you become a developer.
Finding Orbit
This year was a lot to take. Far from home, my stable office job turned chaotic, and layoffs swept away what teams I’d come to respect. On top of that, life had its own challenges to throw my way, while the world of deep learning and software surged forward. Consumers may see only the hype surrounding “intelligent assistants”, but the deeper disruptions continue to reshape the global market.
Staring burnout in the face, and seeing many others experiencing the same thing, gave me a new perspective. I tracked what worked, what didn’t, and where my time was best spent. All of these things forced me to evaluate what I cared about: making things, learning things, and forming connections with people who make the work meaningful.
It’s a bit of a hopeful footnote to the year. Whether the probationary period moves forward or not, it feels like a guidepost has been firmly planted. Imposter syndrome sucks, but every code review, every question I have, every stand-up meeting is a reminder of how much there is left to learn, and I find it motivating. You come to realize most people are in the same boat, and that shared struggle and communication becomes a map; it guides you into accumulating new perspectives, forcing you to poke around problem spaces you would’ve otherwise never been exposed to.
I’ve found a stable enough orbit to keep learning while moving forward for a bit. I like to think that’s enough.
“Our intuition about the future is linear. But the reality of information technology is exponential, and that makes a profound difference. If I take 30 steps linearly, I get to 30. If I take 30 steps exponentially, I get to a billion.”
- Ray Kurzweil When Humans Transcend Biology