Blog
Notes on turning a good idea into a working system.
We write about concrete product, design, and engineering decisions - situations we have encountered in our own work.What do we write about?
- Decisions that genuinely shape a project
- Design and engineering in the same story
- Patterns and lessons from real delivery
Featured article
What is RAG, and how does AI get access to your company's own knowledge?
Language models know everything about the world — and nothing about your company. RAG bridges that gap: before answering, it retrieves the relevant knowledge from your own documents. Here is how it works, what it is for, and what its success depends on.
Read the articleTopics
Choose the area that is most relevant to your work right now.
What is RAG, and how does AI get access to your company's own knowledge?
Language models know everything about the world — and nothing about your company. RAG bridges that gap: before answering, it retrieves the relevant knowledge from your own documents. Here is how it works, what it is for, and what its success depends on.
What is TailwindCSS, and why should you use it?
TailwindCSS is a popular utility-first CSS framework that lets developers shape a website's look quickly and efficiently — building fully custom designs from small, composable classes.
Why choose Next.js for building websites?
Next.js is one of the most advanced React-based frameworks, with real benefits for developers and users alike. We look at what makes it special — and when it is the best choice.
The role of the Top-k parameter in large language models (LLMs)
Top-k limits how many candidate words a language model considers for the next word — balancing creativity and coherence. Here is how it works, with a concrete example.
The role of the Top-p parameter in large language models (LLMs)
Top-p — also known as nucleus sampling — decides how wide a pool of candidate words a model draws from. Here is how it works, and how to configure it well.
The role of the temperature parameter in large language models (LLMs)
The quality and style of AI-generated text depends heavily on parameter settings — and one of the most important is temperature. Here is what it does and how to tune it.