Dylan M.

Key Skills

Software

Other

Internal/Proprietary Tooling

Top Subject Matter

AI Model Evaluation & Validation (LLM / RLHF)

Biological Sciences (Physiology, Toxicology & Neuroscience)

Technical Writing, Scientific Manuscripts & Data Integrity

Top Data Types

Text

Image

Audio

Top Task Types

Evaluation/Rating

Transcription

Object Detection

Text Summarization

Question Answering

Text Generation

Prompt + Response Writing (SFT)

Freelancer Overview

As a Technical Research Fellow and Graduate Biology Researcher, I specialize in the rigorous alignment, validation, and evaluation of Large Language Models (LLMs). My expertise spans advanced Reinforcement Learning from Human Feedback (RLHF) workflows, multi-turn prompt engineering, and truthfulness/factual verification. I bring deep technical knowledge in physiological sciences, toxicology data structures, and complex nomenclature, allowing me to audit model responses for high-level scientific accuracy and adherence to rigid logical schemas. In addition to textual evaluation, I have extensive experience grading complex multimodal data arrays—including image, audio, and video streams—to cultivate high-fidelity training inputs for neural networks. Backed by a strong foundation in statistical software (SPSS and R) and peer-reviewed scientific manuscript preparation, I excel at stress-testing AI logic, eliminating algorithmic hallucinations, and ensuring data integrity across complex technical domains.

Labeling Experience

Technical Research Fellow (AI Evaluation) — Handshake AI (November 2025 – Present)

OtherTextEvaluation/Rating

Role involves evaluating and ranking outputs from Large Language Models (LLMs) to refine model behavior and ensure training-quality responses. You assess complex multimedia content and assign ratings/ordering to support high-fidelity inputs for neural network training. The work includes collaborative testing with frontier AI research labs to improve evaluation rigor and output reliability. • Evaluate and rank LLM outputs via rigorous output assessment. • Rate and order complex multimedia content (image, audio, video) for training suitability. • Contribute to refining LLMs through structured comparisons. • Ensure high-fidelity training inputs for neural networks.

2025 - Present

Education

Work History