This is the third and final post in my mini-series. In my first post, Where have all the Engineers gone?, I shared some hard questions about the current state of engineering and what it will look like a few years from now. In my second post, Rise of the Agent Director, I discuss the importance of experienced senior engineers and their impact in helping to manage and direct coding agents. We ended that with the challenge of what can early-career engineers do now? Let’s explore that here!
As AI becomes more embedded in engineering workflows, there’s a dangerous temptation emerging: we can shortcut real experience. We don’t need to train junior engineers anymore because AI can generate the code. We can downsize teams and rely on automation to fill the gaps. Prompting is enough.
Lies.
Subscribe to my blog and get posts like this in your inbox. Share your email below, or follow me on Threads, LinkedIn, or BlueSky.
“Laying off developers because of AI is the dumbest move in tech right now.” (Perazzo, 2024)
Organizations that take this path risk gutting their engineering pipelines. They risk losing the deep, critical knowledge that only comes from years of trial, error, and mentorship. And worst of all, they might not realize the damage until it’s too late. Systems start to rot, architectural coherence slips, and no one’s left who understands how it all works.
We need to continue to invest in growing engineers, real ones. The kind who think, question, debug, and design. The kind who can lead in a future where AI is everywhere, but judgment still matters most.
AI can write it. They decide whether it’s right.
What might this look like?
Intentional Growth for Junior Engineers
Today’s junior developers need to be more intentional than ever about their growth. It’s incredibly easy to get lazy and let AI do the heavy lifting, but that short-circuits real learning. This isn’t a jr engineer problem, it’s a human problem. Those of us who came up before AI didn’t have to think so hard about being deliberate; we had to troubleshoot, write from scratch, and grind through the process. How many nights I burned chasing down bugs… That’s how the craft was built. Now, that same experience must be pursued with great intentionality.
“AI can assist with code generation, but without mentorship, [our] code quality, security, and architecture suffer.” (Osmani, 2024)
To ensure today’s junior engineers grow into the thoughtful, system-oriented senior engineers we’ll need tomorrow, I encourage some practical techniques:
Engage with AI
AI isn’t going anywhere; it’s too hard to avoid. However, it doesn’t mean you have to give up and let it do everything, not allowing you to learn. It can be of great use and impact to your learning journey!
Shadow the AI. Use AI to generate solutions and code, but take time to understand and question its decisions. Rewrite its code manually to compare approaches. Search the web for concepts it’s using and ensure they’re accurate. Treat the AI like a senior engineer, watch, question, and learn.
“Learning is now on-demand. I don’t need to know Python to build a Hello World app. I can build the app, then understand how it works.” (Palmer, 2025)
Question the AI. Ask the AI the same question more than once! If you get different answers each time, review what makes them different. Try different LLMs entirely, i.e. OpenAI vs Claude, and compare.
Keep an AI contrast journal. Log the AI-provided solutions and comparisons alongside alternatives you’ve researched or come up with. Reflect on which choices are better and why.
Be critical of AI, don’t just assume it’s right (rarely is it that good). Ask questions, make changes, and engage with its outputs.
Code without AI
You don’t have to use AI. Give it a go!
Go old school (are the kids calling it vintage now?). Grab a coding book. Use Google to search, ignore its Gemini response, read blogs and forums, find code snippets, copy/paste, and write your own code!
AI-free problem-solving. Beyond coding, don’t use AI in debugging issues either. Nothing feels better than spending 4 hours troubleshooting a bug when you realize you mistyped a variable.
Own a real component. Maintain responsibility for a small but non-trivial piece of software. Handle its evolution, debugging, and design choices.
This is how our experienced engineers became experienced, building critical thinking and system design skills: hands-on code, learning new patterns through trial and error, and mind-numbing troubleshooting.
Engage with People
Outside of your own learning path, engage with experts who have gone before you.
Read great code. Study well-maintained open source projects to internalize best practices and decision patterns. Use AI to explain certain complexities and ask why the authors might’ve chosen the direction they did.
Seek out human pairing. Participate in reviews with senior engineers, not just for code validation but to understand architectural and strategic thinking. Be curious, ask questions.
Share what you learn. Explain your work to others, how you reasoned through a complex situation. Consider speaking at user groups or blogging. Repeating what you learned to others reinforces these facts for yourself.
Your future is in your hands
This takes time and effort, and chances are, it will happen outside of work. Make coding your new hobby, take on side projects, and be eager to learn! We don’t know the true future of software engineering, even if it is 100% AI-written code in 5 years, what you can learn today, through the above tips, will be invaluable. We’re going to need you, start today.
Engineering leaders, we need to show up
I appreciate it’s one thing to say all of this, and it’s another to enable some to do it. Unfortunately, most engineering teams are driven by deadlines and tight budgets. Letting an engineer learn costs time and money. Is the investment worth it? Yes, I think so!
Organizations must invest now in developing the next generation of “engineers”, or Agent Directors, or whatever clever name someone comes up with. It’s the future, but we need to plan now. Failing to build future senior, experienced engineers introduces long-term gaps in leadership, technical expertise, and innovation.
Subscribe to my blog and get posts like this in your inbox. Share your email below, or follow me on Threads, LinkedIn, or BlueSky.
David Lozzi 