Brennan T.

Key Skills

Software

Other

Mercor

Top Subject Matter

Multimodal AI evaluation and annotation

Multimodal AI evaluation

preference comparison

Top Data Types

Image

Video

Text

Audio

Document

Top Task Types

RLHF

Evaluation/Rating

Transcription

Data Collection

Question Answering

Bounding Box

Freelancer Overview

I have experience in AI training data, data labeling, and multimodal evaluation across image, video, audio, and text-based projects. My work has included evaluating AI-generated outputs for accuracy, instruction-following, visual quality, and consistency, as well as performing detailed annotation and quality assurance tasks for machine learning datasets. I have worked with complex visual reasoning guidelines involving object tracking, referring expressions, segmentation, image comparison, and prompt evaluation, where precision and consistency were critical. I also have experience reviewing model failures, identifying hallucinations, verifying dataset quality, and applying strict annotation standards across large-scale workflows. What sets me apart is my strong attention to detail, ability to quickly learn complex project guidelines, and experience working on high-accuracy evaluation tasks under tight quality standards. I am skilled at identifying ambiguity, ensuring annotation consistency, and providing structured reasoning for decisions. My background includes multimodal AI evaluation, video QA, entity recognition, prompt assessment, and comparative ranking tasks, along with experience using labeling and annotation platforms to support AI model training and reinforcement learning workflows.

Labeling Experience

AI Evaluation & Annotation Specialist (Handshake AI contract)

Image

Evaluate AI-generated multimedia outputs (images, video, and text) for accuracy, instruction adherence, visual consistency, and overall quality against detailed guidelines. Perform entity recognition and prompt verification for celebrities, landmarks, products, fictional characters, brands, and other multimedia entities to support training data reliability. Analyze video scene detection, camera movement, and action segmentation with precise timeline coverage to ensure high-fidelity annotations. • Conduct multimodal quality reviews to ensure outputs meet evolving standards. • Identify issues such as specificity, ambiguity, and information leakage in recognition/knowledge prompts. • Produce structured annotations and verification data for model training, QA, and multimodal evaluation. • Complete high-volume evaluation tasks with strict attention to detail.

2026 - Present

Audio Model Trainer / Digital Annotation Expert (Mercor contract) duplicate summary (subset)

OtherAudio

Create structured training data by evaluating multimodal AI outputs against prompts to ensure model alignment and output quality. Apply detailed annotation guidelines and quality benchmarks to support consistent labeling across large-scale workflows. Provide feedback on observed inconsistencies to support iterative model refinement. • Perform preference-based ranking and comparisons relevant to model alignment. • Manage high-volume annotation tasks with accuracy and speed. • Produce reliable dataset artifacts for model fine-tuning workflows. • Shift between evaluation and production annotation roles as required.

2025 - 2026

Audio Model Trainer / Digital Annotation Expert (Mercor contract)

OtherAudioRLHF

Evaluate and rank multimodal AI outputs (audio, image, and video) against prompts to assess alignment, relevance, and overall output quality. Execute preference-based comparisons using standardized evaluation frameworks to support reinforcement learning from human feedback (RLHF). Perform high-volume digital annotation work following detailed labeling guidelines and quality benchmarks to help improve model performance. • Maintain high accuracy and productivity under tight turnaround requirements. • Provide structured feedback on quality trends and inconsistencies found in model outputs. • Use specialized annotation platforms to produce training datasets for fine-tuning. • Adapt between evaluation-focused and production annotation responsibilities as project needs change.

2025 - 2026

Education

Work History