Things worth
thinking through.

HTML describes what content means. A navigation element contains links for moving around a site. A heading introduces a section. A label tells someone what a form field expects.

CSS decides how those elements look and fit together. The same HTML can become a single column on a phone or a spacious layout on a desktop. Keeping structure and styling separate makes both easier to change.

This site uses five HTML documents and one shared stylesheet. The small amount of JavaScript adds the liquid-metal accent and prepares contact drafts; the pages remain readable without it.

Read the Starter Pack’s semantic HTML lesson ↗

A search engine can find a paper that mentions two concepts. A graph can also represent how those concepts relate. Combining retrieval and graph traversal gives a system more than a list of matching documents to work with.

That does not make an answer automatically correct. The useful part is keeping the evidence visible: what was retrieved, which relationships were followed, and what the sources actually support.

Biomedical GraphRAG is a project exploring this combination of document retrieval and connected information.

See the project ↗

A marketplace is more than a screen of listings. People need to know whether they can trust an account, find what they are looking for, contact someone, and understand what happens next.

Those questions reach across product design and engineering. Authentication, search, moderation, and messaging all shape the experience. Deciding how they fit together is part of building the product, from its earliest scope through its release.

See PolyBuys ↗