Leonard Okinyi

Native English-proficient AI Generalist and Data Trainer with 3+ years of hands-on experience evaluating AI-generated outputs, developing high-quality training data, and providing structured feedback to improve LLM and generative AI model performance. Proven generalist with expertise spanning linguistics, media, technical content, and localization — enabling credible evaluation across diverse tasks and domains. Adept at analyzing AI outputs for clarity, accuracy, and relevance using structured rating frameworks, with a track record of maintaining 98%+ accuracy in high-volume annotation workflows.

Collaborative, adaptable, and deeply curious — ready to contribute to cutting-edge AI agent projects.

Data Annotator & AI Content Quality Evaluator at RWS Group (2024-08 – Present)

• Evaluated AI-generated multimodal outputs across diverse tasks using structured guidelines and predefined rating frameworks, assessing for accuracy, fluency, semantic quality, and cultural appropriateness — directly aligned with AI agent training goals.
• Identified and documented recurring quality gaps and error patterns in AI outputs, producing structured written justifications and actionable feedback that improved NLP and generative model performance.
• Conducted URL grading and multilingual content quality assessments, applying advanced annotation techniques to support model training pipelines at scale.
• Collaborated with evaluation and production teams to refine annotation methods and develop linguistic and content guidelines for ongoing AI model improvement.
• Created culturally appropriate prompts for AI-driven content tools, leveraging generalist knowledge to ensure outputs met brand voice and communication standards across diverse domains.

AI Technology & Localization Consultant at Appen (Present)

• Fine-tuned generative AI models for multimodal and multilingual applications, evaluating text and image outputs for realism, accuracy, and production-level quality across a broad range of subject areas.
• Applied evaluation rubrics to assess AI-generated content, flagging quality gaps related to clarity, composition, and production standards — bringing generalist cross-domain judgment to each task.
• Maintained a 98%+ accuracy rate across complex content review and annotation workflows, demonstrating exceptional consistency and attention to detail in high-volume environments.
• Collaborated with designers and creative teams to localize and adapt marketing, UI, and media content for target markets, ensuring visual and linguistic consistency.
• Managed translation memories and content style guides using Phrase, Smartling, and Trados to maintain consistency across large-scale multilingual projects.

AI Data Trainer & Media Content Reviewer at Upwork (Contract) (2023 – 2023)

• Managed content review and validation projects for software, website, and media localization, performing structured linguistic quality testing across multiple formats and industries.
• Conducted structured quality assessments on AI-generated content, evaluating accuracy, fluency, and appropriateness against client specifications across diverse subject domains.
• Collaborated with cross-functional teams to document evaluation criteria and refine content workflows for AI training data creation.
• Implemented quality-check workflows for digital media, optimizing for global platforms and ensuring adherence to production and visual standards.

Patent & Technical Content Reviewer at Independent / Various Clients (Present)

• Translated and validated high-stakes technical documentation, applying rigorous structured review to ensure accuracy and consistency across complex, specialized content.
• Developed and maintained specialized glossaries and evaluation criteria for technical and legal content — demonstrating the kind of domain-switching and interdisciplinary adaptability this role demands.

Bachelor of Arts in Linguistics – University of Nairobi, Kenya