Peter Kigen

Performed high-volume data annotation, content labeling, and quality review for large language model training datasets. Conducted preference ranking (RLHF-style) of model responses and maintained consistent quality scores above platform averages.

Executed text classification, entity recognition, sentiment analysis, and content moderation tasks for computer vision and NLP model training. Built strong foundation in following precise labeling guidelines and delivering consistent, high-accuracy annotations under volume targets.

Reviewed and rated AI-generated text for factual accuracy, relevance, tone, and policy compliance as part of large-scale RLHF and supervised fine-tuning data pipelines. Applied complex multi-dimensional rubrics to score responses and provided structured written feedback.

Activities and societies: Capstone Project: Developed and evaluated a multi-turn conversational AI agent focused on factual accuracy and safety alignment; conducted systematic response ranking and preference modeling (RLHF-style evaluation) across 1,200+ generated outputs.

Bachelor of Science in Artificial Intelligence with relevant coursework in machine learning, natural language processing, AI ethics and safety, data mining, human-computer interaction, and statistical analysis. Capstone project involved developing and evaluating a multi-turn conversational AI agent focused on factual accuracy and safety alignment.