santiago lencina

I evaluate AI-generated Spanish responses for factual consistency, instruction adherence, tone, clarity, and contextual appropriateness, with particular attention to natural Rioplatense Argentine phrasing.

Through independent AI evaluation, annotation, and prompt-testing projects, I create and refine prompts, compare outputs across AI systems, identify failure patterns, and rewrite responses for greater precision and usefulness.

I apply explicit rubrics to annotation-style work including entity identification, response assessment, written rationales, and quality review. I document reproducible issues, isolate problematic wording or behavior, and propose clear corrections.

I also build and test React and TypeScript web applications, including dashboards, service-management tools, and data-driven interfaces. My technical QA and local LLM workflow experience helps me troubleshoot defects, validate fixes, and communicate issues clearly.