Looks like you have JavaScript disabled. For the full Twine experience, you will need to re-enable it.

RLHF Specialist and AI Trainer with 6+ years of experience shaping large language model behavior through structured preference data, response evaluation, and adversarial testing. I build and refine annotation guidelines, author prompts and reference answers, and calibrate evaluations for consistency and safety alignment. I’m trusted across multiple platforms and model types for accurate, repeatable labeling and rapid onboarding to new guidelines. From red-teaming and inter-annotator agreement to code and factual verification, I help teams improve model quality with rigorous, guideline-driven QA.…RLHF Specialist and AI Trainer with 6+ years of experience shaping large language model behavior through structured preference data, response evaluation, and adversarial testing. I build and refine annotation guidelines, author prompts and reference answers, and calibrate evaluations for consistency and safety alignment. I’m trusted across multiple platforms and model types for accurate, repeatable labeling and rapid onboarding to new guidelines. From red-teaming and inter-annotator agreement to code and factual verification, I help teams improve model quality with rigorous, guideline-driven QA.

Michael Tyree

AI Trainer, AI Evaluator, Data Annotator, +1





RLHF Specialist and AI Trainer with 6+ years of experience shaping large language model behavior through structured preference data, response evaluation, and adversarial testing. I build and refine annotation guidelines, author prompts and reference answers, and calibrate evaluations for consistency and safety alignment. I’m trusted across multiple platforms and model types for accurate, repeatable labeling and rapid onboarding to new guidelines. From red-teaming and inter-annotator agreement to code and factual verification, I help teams improve model quality with rigorous, guideline-driven QA.…RLHF Specialist and AI Trainer with 6+ years of experience shaping large language model behavior through structured preference data, response evaluation, and adversarial testing. I build and refine annotation guidelines, author prompts and reference answers, and calibrate evaluations for consistency and safety alignment. I’m trusted across multiple platforms and model types for accurate, repeatable labeling and rapid onboarding to new guidelines. From red-teaming and inter-annotator agreement to code and factual verification, I help teams improve model quality with rigorous, guideline-driven QA.

Available to hire

RLHF Specialist and AI Trainer with 6+ years of experience shaping large language model behavior through structured preference data, response evaluation, and adversarial testing. I build and refine annotation guidelines, author prompts and reference answers, and calibrate evaluations for consistency and safety alignment.

I’m trusted across multiple platforms and model types for accurate, repeatable labeling and rapid onboarding to new guidelines. From red-teaming and inter-annotator agreement to code and factual verification, I help teams improve model quality with rigorous, guideline-driven QA.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Intermediate

Work Experience

AI Trainer / Data Annotation Specialist at Freelance (Remote)

January 1, 2019 - Present

Collected and labeled preference data across text, code, and conversational domains to support RLHF and supervised fine-tuning pipelines. Ranked and evaluated AI-generated responses for accuracy, helpfulness, and safety alignment using detailed project guidelines. Conducted red-teaming of model outputs to surface policy violations, factual errors, and reasoning failures across multiple LLM projects. Authored and edited prompts, reference answers, and ground-truth content incorporated into model training and evaluation datasets. Leveraged Python, Java, and SQL to review, debug, and assess code snippets across several programming languages. Maintained high accuracy and inter-annotator agreement while working independently under shifting guidelines and tight deadlines.