Daniel R.

Key Skills

Software

Other

Top Subject Matter

AI safety

frontier model evaluation

adversarial prompting

Top Data Types

Text

Video

Image

Top Task Types

Red Teaming

Freelancer Overview

AI Safety & Model Evaluation (frontier model testing, user simulation, rubric-based assessment, reasoning analysis, adve. Brings 11+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Other. AI-training focus includes data types such as Text, Video, and Image and labeling workflows including Evaluation, Rating, and Red Teaming.

Labeling Experience

100+ hours of multimodal AI evaluation across text, image, video, and audio systems

OtherImage

Performed evaluation-focused annotation work for multimodal systems by rating and analyzing model outputs for quality and behavioral alignment. Used structured review and reasoning analysis to interpret responses in a rubric-guided manner. This included multimodal evaluation across text, image, video, and audio content. • Rated multimodal outputs against safety/quality criteria • Performed structured quality review and reasoning checks • Used consistent rubric-based methods during testing • Analyzed behavioral elicitation results across modalities

2024 - Present

Independent multimodal AI evaluation for safety research (adversarial scenario testing and failure-mode analysis)

OtherVideoRed Teaming

Conducted red-teaming style assessment activities by generating adversarial interactions and systematically probing model failure modes across modalities. Applied structured evaluation methods to characterize behavioral weaknesses and safety-relevant outputs during testing. These efforts supported independent research into alignment and language-mediated control surfaces in AI systems. • Probed model behavior under adversarial conversational conditions • Assessed quality using consistent evaluation rubrics • Analyzed failure modes and reasoning characteristics • Extended evaluation coverage across multimodal systems

2024 - Present

AI Safety & Model Evaluation (frontier model testing, user simulation, rubric-based assessment, reasoning analysis, adversarial conversational scenarios)

OtherText

Performed multi-turn conversational testing and behavioral elicitation to evaluate frontier model responses using rubric-based and structured quality assessments. Designed adversarial prompting scenarios to surface targeted behaviors and failure modes, then analyzed reasoning quality and overall response performance. Completed 100+ hours of multimodal evaluation work spanning multiple content types, with a focus on elicited model behavior and alignment-relevant characteristics. • Used persona-driven user simulation to elicit specific behaviors • Applied rubric-based scoring and written model analysis • Conducted reasoning evaluation and structured quality review • Selected for advanced reviewer-track responsibilities based on performance

2024 - Present

Education

Work History