Articles

Writing about software engineering, machine learning, and the tools I use every day. Deep dives, tutorials, and lessons learned from building real-world applications.

Sort by
ZSOM: A Production-Ready Self-Organizing Map in Pure NumPy
Jun 13, 202650 min read

ZSOM: A Production-Ready Self-Organizing Map in Pure NumPy

A deep dive into ZSOM — a full-featured, Numba-accelerated SOM implementation with 3D mesh support, adaptive learning, hex topology, and publication-quality visualization. Covers the math behind every design decision.

Building This Portfolio: Vike, an In-Build SQLite, and Two Content Engines
Jun 12, 202616 min read

Building This Portfolio: Vike, an In-Build SQLite, and Two Content Engines

How this portfolio runs a fully searchable, bilingual, content-rich site with no backend — just a SQLite database generated at build time and shipped inside the bundle. A feature-by-feature tour, plus the two open-source content engines (HyperDown + HyperJson) you can reuse to build the same thing.

What is HyperDown? Markdown → SQLite Full-Text Search, SSR-Only
Jun 6, 202612 min read

What is HyperDown? Markdown → SQLite Full-Text Search, SSR-Only

Add real full-text search to your site with no backend, no API, and no CMS server. HyperDown turns a folder of Markdown/MDX into a typed, searchable SQLite layer that ships inside your build — here is why you'll want it and how to start in minutes.

What is HyperJson? Schema-First JSON Content for Vite
May 30, 202610 min read

What is HyperJson? Schema-First JSON Content for Vite

Stop hand-writing TypeScript interfaces for your JSON content. HyperJson validates every JSON file against a schema at build time and generates the types for you — invalid content fails the build, valid content arrives fully typed, with no backend and no runtime cost.