Bruce O.

Key Skills

Freelancer Overview

Over the past few years, I've been involved in several AI training and data labeling projects across different modalities, including text classification, named entity recognition (NER), and sentiment analysis. I've worked with annotation platforms like Scale AI and Labelbox, where I was responsible for tagging datasets according to detailed style guides and ontologies. I'm comfortable handling edge cases, escalating ambiguous labels, and maintaining consistency across large annotation batches. I also have experience with inter-annotator agreement (IAA) scoring and understand how disagreements are resolved to improve label quality. Beyond basic labeling, I've contributed to RLHF (Reinforcement Learning from Human Feedback) tasks — specifically writing and ranking model responses to help fine-tune conversational AI outputs. I understand the importance of preference data, how reward models are trained, and why calibration between annotators matters for downstream model performance. I'm familiar with prompt engineering principles and have applied them while evaluating model outputs for helpfulness, harmlessness, and factual accuracy. I'm confident working within annotation pipelines that require both speed and precision, and I know how to balance throughput with quality to meet project SLAs.

Education