Technologies

Prompt engineering

Designing and combining prompts and generation parameters so that small language models produce useful text.

• Python

Small language models

Using small-parameter, freely available language models on minimal hardware.

• Python

Template-based document generation

Separating content (structured data) from presentation (templates), so one build step produces every version of a document from a single source.

• Python

Static site generation

Building the HTML of websites from templates and data at build time, so the server only serves finished files. Here extended to several domains from one source.

• Python

Incremental deploy

Uploading only the files that changed. The key question is change detection: FTP offers file size and modification time, but no content hash.

• Python

Web performance

Faster loading and a stable layout, measured with Core Web Vitals. On image-heavy static sites, image handling matters most.

• Python

Technical SEO

The technical structure of a site that affects crawling, indexing and how it appears in search results, separate from writing content and building links.

Privacy-friendly web

Web pages that send no visitor data to third parties by default: every front-end dependency is served from the site's own hosting.

Content collections

Structured content in files, one collection per type, validated by a schema at build time, so pages read typed data instead of hard-coded values.

• TypeScript

Bilingual content model

Every text in every language lives side by side in one source with the same structure, and the language is part of the URL.

• TypeScript

Client-side search

Full-text search with an index built as a static file at build time and queried in the browser, without a server or a third-party service.

• TypeScript

Knowledge graph visualization

An interactive force-directed graph generated from content relations that shows how languages, technologies, projects and research connect, without levels or scores.

• TypeScript