Thomas Pentzek

I’m an AI evaluator and content writer specializing in reviewing and improving AI/LLM outputs for relevance, correctness, completeness, tone, and guideline/policy compliance. I bring a strong quality and process mindset—focused on turning model mistakes into clear, actionable improvements.

In Outlier projects, I’ve actively worked as an AI Evaluator/AI Rater for years, scoring and ranking responses using rubrics across correctness, relevance, completeness, tone, and safety. I write precise rationales and improvement suggestions to support model optimization, and I consistently identify recurring error patterns such as hallucinations, incorrect conclusions, and missing context.

I also have 10+ years of professional content writing experience, creating and revising SEO and editorial content like web articles, guides, explainers, and content updates. I execute client briefs with clear structure and high text quality, working with CMS platforms (e.g., WordPress) and applying basic SEO practices to keep content helpful and discoverable.

Before that, I led international teams as a Plant Manager in Mexico, where I managed production performance, quality, delivery KPIs, and continuous improvement. That background strengthened my ability to document edge cases, manage feedback loops, and maintain consistent standards—skills I apply directly to AI/LLM quality assurance and structured writing.