About Me: Building, Breaking and Learning Through Software

Hi, I’m Marceli Pawliński.

I’m a student from Scotland interested in computer science, cybersecurity, artificial intelligence and software engineering.

Most of what I know has come from building things, breaking them, fixing them, and then realising there was probably a better way to build them in the first place.

How I Got Into Computing

Computers have been part of my life for as long as I can remember.

Over time, that turned from simply using technology into wanting to understand what was happening underneath it.

I started experimenting with programming, websites, servers, local AI models and cybersecurity.

The more I learned, the more interested I became in what happens when systems stop behaving the way they are supposed to.

That is probably the part of computing I enjoy most.

A system working correctly is useful.

A system failing in an unexpected way is interesting.

What I Build

My main project at the moment is Zenith.

Zenith is a platform designed around one idea:

Monitor, improve and grow your website.

It brings together website monitoring, security checks, performance analysis, accessibility, SEO and growth opportunities.

Building Zenith has pushed me into areas I probably would not have explored otherwise, including:

• Cloudflare Workers
• databases
• authentication
• queues
• APIs
• scheduled jobs
• browser automation
• security headers
• rate limiting
• website auditing
• AI-assisted analysis

It has also taught me that building a product is very different from building a feature.

A feature can work perfectly on its own and still make no sense inside the wider system.

Open Source

I also spend a lot of time on GitHub.

My GitHub is a mixture of personal projects, experiments and contributions to other repositories.

One project I’ve been working on is browser-harness, where I’ve been exploring how AI systems can interact more reliably with websites and browser interfaces.

I also enjoy contributing to existing open-source projects.

Working on someone else’s code is very different from starting from an empty file.

You have to understand why something was built a certain way before deciding how to change it.

Sometimes a useful contribution is only a tiny fix.

That is still valuable.

AI and Local Models

I started properly experimenting with large language models around 2023.

Since then I’ve worked with tools and models including:

• Ollama
• llama.cpp
• LM Studio
• Qwen
• Llama
• DeepSeek

At first I was mainly interested in the models themselves.

Eventually I became more interested in everything around them.

How does an AI use tools?

How do you give it permissions safely?

How does it interact with browsers?

How do you handle memory?

How do you stop an automated system from doing something it shouldn’t?

How do you make it reliable when the real world is unpredictable?

Those questions are much more interesting to me than simply asking which model has the highest benchmark score.

Cybersecurity

Cybersecurity is another major interest of mine.

I especially like looking at the difference between how secure something appears and how secure it actually is.

A polished login page does not necessarily mean the backend is secure.

A website having HTTPS does not mean everything behind it is configured correctly.

A security feature existing does not mean it has been implemented properly.

That way of thinking has changed how I approach software development in general.

Instead of only asking:

“Does this work?”

I also try to ask:

“How could this fail?”

Learning Outside School

I’m currently studying subjects including Maths and Computer Science, but a lot of my technical learning happens outside the classroom.

I experiment with my own infrastructure, contribute to GitHub projects, build websites, run local AI models and work through technical certifications.

I’ve also completed learning in areas including AI literacy and introductory cybersecurity.

I like structured learning, but I learn best when I can actually use something.

Reading about databases is useful.

Designing one badly, discovering the problems and then fixing it teaches you something different.

What I Want to Do

Long term, I want to work somewhere around the intersection of:

AI, cybersecurity, systems and computer science.

I’m still figuring out exactly what that looks like.

That is partly why I build so many different things.

Every project gives me a slightly better idea of what I enjoy, what I’m good at, and what I still need to understand.

Right now, I’m focused on improving Zenith, contributing more to open source and continuing to build projects that force me to learn things I don’t already know.

If something breaks along the way, that normally makes it more interesting.

---

Find me

GitHub: github.com/marceli1404

Website: marcelipawlinski.co.uk

Zenith: zenith.me.uk

I’ll be writing here about software development, cybersecurity, AI, open source, website infrastructure and whatever I happen to be experimenting with next.