Owen T.

Cognitive Scientist and Multimodal AI Trainer

Chapel Hill, Usa

Key Skills

Software

No software listed

Top Subject Matter

My strongest subject matter areas are cognitive neuroscience

behavioral psychology

machine learning. Through my academic work at UNC

predictive modeling research in the DEPENd Lab

I developed expertise in human decision-making processes

Top Data Types

Image

Video

Text

Top Task Types

Data Collection

Transcription

Point/Key Point

Red Teaming

Polyline

Entity (NER) Classification

Freelancer Overview

My AI training and data labeling experience consists of engineering prompt architectures, curating alignment datasets, and evaluating model performance across Handshake AI, Mercor, and Data Annotation Tech. I have scaled RLHF (Reinforcement Learning from Human Feedback), RLAIF, and DPO (Direct Preference Optimization) training pipelines for frontier generative models. My core work includes designing labeling taxonomies, conducting adversarial red-teaming and prompt-injection testing, and analyzing model outputs to enforce safety guidelines, logical reasoning, and technical instruction-following. I specialize in multimodal data annotation, where I design grounding benchmarks and evaluate vision-language models on spatial semantics, geometric relationships, and image-to-text alignment. I also author technical datasets containing API integrations and algorithmic coding challenges to train and evaluate code-generation models. This data operations work is directly informed by my academic background in computational neuroscience and psychology, where I modeled human decision-making heuristics and analyzed eye-tracking and physiological datasets.

Labeling Experience

My AI training and data labeling experience is built on a strong technical foundation of graduate coursework and hands-o

My AI training and data labeling experience is built on a strong technical foundation of graduate coursework and hands-on laboratory research. In Dr. Michael Hallquist's DEPENd Lab at UNC, I applied machine learning and computational models to generate neuroscience predictions from high-dimensional physiological and neuroimaging datasets. This research is supported by my graduate-level machine learning coursework under Dr. Zachary Fisher in PSYC 559. I translate this theoretical and statistical background directly into my professional alignment work at Handshake AI, Mercor, and Data Annotation Tech, where I design advanced taxonomy schemas, curate RLHF/DPO preference datasets, and benchmark frontier models on complex logical and multimodal reasoning.

Not specified

Education

University of North Carolina at Chapel Hill BS Neuroscience University of Michigan Ann Arbor School of Dentistry DDS

Degree not specified

Not specified

Work History

Mercor

I worked

Location not specified

Not specified