Michael Keraa

I am a Computer Science undergraduate with hands-on experience in AI data annotation and training large language models. I specialize in instruction-following evaluation, response ranking, preference labeling, and factuality verification to improve model behavior.

In my current remote contract role I annotate and evaluate LLM-generated responses against employer rubrics for helpfulness, correctness, instruction adherence, and safety. I review multi-turn conversations, detect hallucinations and logical errors, and escalate ambiguous edge cases with clear written rationale.

I apply task-specific guidelines consistently while meeting quality and throughput benchmarks in asynchronous, high-volume environments. I am familiar with RLHF pipelines, human feedback annotation, and model output review, and I bring strong reading comprehension and written reasoning skills to every task.

I learn onboarding materials quickly, prioritize data quality and AI safety principles, and work well independently or within distributed teams using tools like Slack and Jira. I am pursuing a BSc in Computer Science and seek opportunities to deepen my LLM evaluation and model-alignment expertise.