Work / Language Evaluation
Language data, guideline-based evaluation, classification, documentation, and quality work.
Scope
The work turns guideline-based judgment into repeatable structures without pretending that every language case can be made clean or context-free.
Context
Spanish transcription, web search evaluation, and social media labeling establish the professional history behind the current language-evaluation and classification methods.
language data → evaluation → classification → documentation → quality review
Objects
Published objects connected to language evaluation, classification, and AI education.
Curation schema
Each entry separates meaning in use, register, social context, translation, source, NLP relevance, and claim status. Verified, plausible, and uncertain material remain visibly distinct.
INSPECT THE ARCHIVE →Evaluation graph
A Spanish-language graph for comparing model use by task, interaction dynamic, failure risk, prompting mode, verification, and practical platform fit—rather than relying on a universal ranking.
OPEN THE MAP →Teaching system
A structured public workshop sequence connecting model concepts, prompts, specification, verification, situated context, and practical exercises.
OPEN THE WORKSHOP →Principle
The cleanest category is not always the most truthful one.
A responsible editorial or taxonomy system should show the rule being applied, the source of the judgment, the level of certainty, and the point at which contextual interpretation begins. Ambiguous cases are documentation work—not debris to hide.
CONTACT →