Patrick Losieku

I’m a detail-oriented AI Trainer with two years of experience training and improving generative AI models through prompt writing, response evaluation, and content moderation. I assess AI-generated content for accuracy, relevance, safety, and overall quality while strictly following detailed guidelines. In my work, I created and refined prompts to improve responses, then evaluated outputs using rubrics for consistency—covering accuracy, relevance, clarity, safety, and guideline compliance. I also moderate content by identifying and flagging policy-violating material, and I conduct research to verify factual information before sharing detailed feedback to help improve model performance. I bring strong analytical, research, and English written communication skills, with a proven ability to deliver high-quality work in remote environments. Alongside my background in Criminal Justice and Security Studies and additional qualifications in Cyber Security, I’ve built a careful, confidentiality-minded approach—from managing legal documentation during my attachment to maintaining the same rigor in AI quality assurance.