I'm 23, have a computer science degree, and have been working as a backend engineer for about eight months. My university education covered programming, algorithms, networking, databases, and other fundamentals, but it didn't teach me how a production system works end to end. I had little experience building, testing, deploying, and operating a real application before starting my job.
I now work with Python, FastAPI, JavaScript, SQL, and Azure on software used by real customers. I handle requirements, investigate existing behavior, modify APIs and backend logic, work with databases, occasionally touch the frontend, write tests, open pull requests, respond to reviews, and communicate with other teams.
The problem is that AI is involved in most of my technical process. I use it to locate relevant code, understand architecture, decide where changes belong, implement features, debug, write tests, and review solutions. I can usually understand the code and business context once someone explains them, but if I had to complete a normal ticket without AI, I'm afraid I would struggle to know where to begin.
I often don't know which repository or file to inspect first, how to trace a request from the browser through an API and services to a database, or which hypothesis to test when something breaks. AI helps me get unstuck, but I sometimes feel like I learn what happened only after the solution already exists.
I'm also stuck in a cycle of creating ambitious learning plans. I decide to master JavaScript, SQL, Docker, Linux, system design, cloud, and testing, then buy books, start courses, build roadmaps, and repeatedly search for a better resource. After a few days I become overwhelmed, lose the structure, and abandon the plan before starting another one later.
I know I'm already contributing to a real product, but I want to become an autonomous junior engineer. I want to take a backend ticket, know how to investigate it, form my own hypotheses, implement and test a solution, debug it, and explain why it works. I still want to use AI, but I don't want it to hold my entire engineering process together.
If you were advising an honest 23-year-old junior engineer, what would you suggest focusing on during the next year? And if you once felt that experienced engineers knew an impossible amount, when did that feeling begin to change?
4 Answers
The feeling that experienced engineers know everything is misleading. They’ve encountered many similar failures and systems over years, so they recognize patterns quickly. You build that internal map by repeatedly investigating real problems, reading existing code, observing production behavior, and learning from reviews. A small end-to-end project can help too: build a modest FastAPI application with a simple frontend, database, tests, and deployment, while keeping AI limited to boilerplate and targeted questions.
There’s no shortcut around learning the codebase through direct experience. Start each task without AI, even if your first attempt is imperfect. After making your own plan, use AI in small, focused chunks and verify everything it suggests. Once the change is complete, explain the solution in your own words. The goal isn’t to avoid assistance; it’s to make sure you understand and can defend the result.
Use AI as an amplifier, not as the person driving. Before asking it for a solution, spend 20–30 minutes trying to locate the relevant code and understand the problem yourself. For each ticket, identify the likely entry point, trace the request through the service and database, and write down a few hypotheses. Then ask AI to critique your plan or explain a small section instead of inventing the whole approach. That repeated practice builds the instinct for where to look.
AI isn’t inherently a problem—professional engineers have always used books, documentation, search, and advice from other people. The important distinction is whether you understand the architecture, tradeoffs, and final code. Learn to give AI precise context, watch for hallucinations, and review its output carefully. At the same time, spend enough time designing systems and debugging independently that you remain capable of making technical decisions without it.
Related Questions
How To: Running Codex CLI on Windows with Azure OpenAI
Set Wordpress Featured Image Using Javascript
How To Fix PHP Random Being The Same
Why no WebP Support with Wordpress
Replace Wordpress Cron With Linux Cron
Customize Yoast Canonical URL Programmatically