Promise-scoring model
AI ExploringTesting whether an LLM can consistently classify a political statement as a measurable promise, a vague intention, or neither — the classification step NetaTrack depends on.
Lab
Prototypes, AI experiments and small tools. Some of these will grow into projects, some will be abandoned, and both outcomes are useful.
Testing whether an LLM can consistently classify a political statement as a measurable promise, a vague intention, or neither — the classification step NetaTrack depends on.
A small utility for cleaning mixed Devanagari/Romanised Nepali input before it reaches a database, so search and matching behave predictably.
The design system powering this site: custom properties, fluid type scale and a dark/light theme, with no bundler and no runtime dependency.
A repeatable checklist for shipping PHP projects to shared hosting — SSL, PHP version, permissions, sitemap submission and backups.
Something here worth talking about? Send me a note