I'm a backend software engineer with deep roots in the C# / .NET ecosystem — and in 2025, I made the leap into AI-augmented development. I now use AI as a core part of my engineering workflow: from architectural reasoning and code generation to debugging and spec drafting.
The fundamentals haven't changed — clean architecture, solid domain modeling, scalable systems. But the speed at which I move has. AI doesn't replace engineering judgment; it amplifies it.
"Big change is due to the smallest impacts."
| Layer | Tools & Methods |
|---|---|
| AI Collaboration | Claude, ChatGPT, GitHub Copilot |
| Prompt Engineering | Spec-driven prompting, context engineering, chain-of-thought |
| Code Generation | AI-assisted TDD, scaffolding via LLMs, agent workflows |
| Architecture | Clean Architecture, DDD, CQRS, Microservices |
| Review & QA | AI-assisted code review, automated test generation |
My daily AI workflow follows a simple loop: Specify → Generate → Review → Refine
- Spec-first prompting — I write precise, structured prompts that mirror software specifications. Garbage in, garbage out still applies.
- LLMs for architectural reasoning — When I'm designing a system, I use AI as a thinking partner to stress-test decisions before writing a single line.
- AI-assisted TDD — Generate unit test scaffolds from specs, then implement to make them pass. Faster red-green-refactor cycles.
- Context engineering — Managing token windows, injecting domain knowledge, and chaining prompts effectively is a craft in itself.
The engineers who will thrive in 2026 are the ones who deeply understand what they're building — because you can't direct an AI toward something you can't define yourself.
Currently learning: Life — one commit at a time.