DEV Community: Angelo G.

Without AI Models, Who Am I?

Angelo G. — Fri, 04 Sep 2026 10:43:06 +0000

I graduated in July 2025 with a degree in Computer Engineering. Not long after graduating, I got my first role in the tech industry.

It wasn’t exactly a Software Developer or Software Engineer position. I joined a startup as Technical Support, but the role ended up being a mix of different responsibilities. I worked on software development tasks, automation, deployments, DevOps, troubleshooting, documentation, and other technical work depending on what the team needed.

For someone just starting out, I learned a lot.

But now, I’m in the last month of my contract, and there’s something that has been bothering me.

I’m not completely confident in my skills.

During my time working, I became heavily reliant on AI tools like ChatGPT and Claude. I use them to understand code, debug problems, plan implementations, write code, review my work, and sometimes even figure out what I should do next.

AI helped me accomplish things that I probably wouldn’t have been able to do as quickly on my own.

But that also created a different problem.

Sometimes I ask myself: How much of this can I actually do without AI?

I can work with an existing codebase. I can deploy applications, troubleshoot issues, work with APIs, databases, Docker, servers, automation workflows, and different frameworks. I’ve been exposed to real projects and real problems.

But if you ask me to build something from scratch without AI helping me along the way, I’m not sure how confident I would be.

And now that I’m preparing to apply for my next job, that worries me.

I’m afraid that I became good at working with AI without becoming good enough at the fundamentals behind the things I was building.

So I started thinking about going back to the basics.

At first, it felt like I would be wasting time. I already spent a year working in the industry. Should I really go back and study programming fundamentals again? Should I go back to data structures, algorithms, databases, networking, JavaScript, backend concepts, and all the things I was supposed to understand before worrying about frameworks and tools?

But maybe going back to the basics now is different.

When I was studying these things before, most of them were just concepts to me. I didn’t always understand why they mattered.

Now I have checkpoints.

I’ve seen APIs fail. I’ve dealt with database issues. I’ve worked with authentication, deployments, environment variables, containers, servers, automation, and existing codebases.

Maybe learning the fundamentals again after experiencing these things will make them easier to understand.

Instead of learning something and wondering, “When am I ever going to use this?”, I can connect it to problems I’ve already encountered.

I also don’t think the answer is to stop using AI completely.

AI is already part of how I work, and I think knowing how to use it effectively is a useful skill. But I want to change my relationship with it.

I don’t want AI to think for me.

I want it to help me think.

I want to reach the point where I can look at AI generated code and understand why it works, recognize when something is wrong, explain the decisions behind it, and build the important parts myself when needed.

So instead of giving up on this industry because I feel behind, I want to use that feeling as a reason to get better.

I have real experience now. I know more about what I don’t know. And I think that gives me a better direction than when I was starting from zero.

The question now is where should I start?

Should I pick one language and rebuild my fundamentals from there?

Should I focus on data structures and algorithms?

Should I build projects without AI?

Should I continue using AI, but force myself to understand and explain every piece of code before accepting it?

I’m still figuring that part out.

For developers who have been in the industry longer, especially those who started their careers around the time AI coding tools became common, I’d genuinely like to hear your perspective.

If you were in this position, with some real world experience but weak confidence in your fundamentals, what would you focus on next?

I don’t want to restart my career.

I want to rebuild the foundation underneath it.

DevFlow [ A CLI Tool to End Your Repetitive Project Setup ]

Angelo G. — Sun, 08 Feb 2026 08:45:34 +0000

This is a submission for the GitHub Copilot CLI Challenge

What I Built

I built DevFlow, a Node.js command-line (CLI) tool designed to help developers quickly start and manage their development sessions with a single command.
As a junior developer with some real-world experience, I often get assigned to assist on multiple projects at the same time, fixing minor bugs, applying small enhancements, reviewing code, or helping maintain existing features. Every time I switch projects, I end up repeating the same setup steps:

Navigating to the project directory
Opening the editor
Installing dependencies
Running dev servers
Switching Git branches
Setting environment variables

When this happens across multiple projects in a day, the repetition becomes frustrating and time-consuming.

DevFlow was inspired by that pain point.

Demo

DevFlow

After setting up DevFlow once, starting your projects is really simple. Just save a session and you’re good to go.

Create a new session for a project called crm-integration:

 $ devflow init crm-integration

Now that the session exists, start it:

 $ devflow start crm-integration

sample session running on background

To see all saved sessions:

 $ devflow list

DevFlow Maintenance

helps keep your DevFlow environment healthy

 $ devflow doctor

DevFlow Analytics

Great for seeing where your time goes

 $ devflow stats

DevFlow Settings

Configure global preferences

More Commands
DevFlow also supports more commands:

$ devflow --help

My Experience with GitHub Copilot CLI

GitHub Copilot CLI played a huge role in speeding up my development process.

I used it to:

Generate command logic and CLI structure
Refine shell commands and Node.js scripts
Improve error handling and edge cases
Quickly prototype features before refining them manually

As a junior developer, Copilot CLI felt like having a senior developer beside me—helping validate ideas, suggest improvements, and reduce time spent on boilerplate code. It allowed me to focus more on designing the workflow and user experience rather than getting stuck on repetitive implementation details.

Overall, Copilot CLI helped me move faster, learn better patterns, and confidently complete this project.

DevFlow is a reflection of my real experience as a junior developer trying to be more efficient while juggling multiple responsibilities.

This project may be simple, but it solves a very real problem I personally face—and I plan to continue improving it over time.