Turning AI from impressive demos into systems you can truly rely on.
I'm Noxtrel, an independent developer who enjoys exploring technology, taking complex problems apart, and turning ideas into things that work.
I focus on where large language models meet the real world: agent architecture, context engineering, RAG, voice interfaces, automation, and system design. I care less about dazzling demos and more about whether a system runs reliably, keeps improving, and actually solves the problem.
This is where I write about my projects, technical experiments, the mistakes I learn from, and my curiosity about technology and the world.
What I work on
Context engineering
Give the model the right information at the right time instead of stuffing everything in.
Agent engineering
Tool use, task loops, harnesses, and evals that make agents controllable and reproducible.
Shipping AI
From requirements and model selection to launch and acceptance, delivered with engineering discipline.
Latest writing
The first essay is on its way.
Course
The Little Token Seller: AI Collaboration & Agent Engineering
A structured course for tech professionals: from how LLMs work, prompts and context engineering, to agents, harnesses, AI project delivery, and two complete hands-on projects.
- 14 modules
- 124 lessons
- 2 end-to-end projects
- Side-by-side examples in every lesson