Derrick Denkyi

I’m a results-driven AI Training Data Specialist and expert evaluator with close to a decade of progressive experience across client communications, data analysis, AI model evaluation, and information technology. I’m known for handling interactions with clarity, professionalism, and genuine care—while bringing structure, ownership, and a focus on measurable outcomes.

On the Mindrift platform by Toloka, I evaluate and rate AI-generated responses for conversation quality, instruction-following accuracy, factual correctness, reasoning coherence, and safety. I deliver expert-level assessments that directly shape the behavior and improvement of large language models, often working across multiple projects with strict professional standards and disciplined evaluation criteria.

Previously, I led email operations and client communications by designing structured inbox triage, achieving 100% order delivery accuracy through meticulous verification, and resolving recurring issues at root cause. I also produced clear, data-driven reports for multiple departments and maintained accurate digital records through consistent QA and process improvement.

My foundation in ICT and AI fluency—supported by Anthropic certifications across AI fluency, developer/API skills, and agentic workflows—keeps my evaluations grounded in how AI systems think and where they fail. I bring prompt engineering, AI alignment mindset, and practical operational discipline to every task I complete.