AI QUALITY ANALYTICS
DETAILED EVALUATION

AUDITING THE

INTELLIGENCE.

PROFESSIONAL EXPERIENCE

Handshake AI

AI Quality Analyst / LLM Evaluation Specialist
OCT 2024 — PRESENT Lead Quality Reviewer
CORE_SKILLS
NVIDIA OPENAI RLHF QA of QAs
English-Domain RLHF Auditing: Acted as the QA of QAs; auditing primary reviewer outputs and adjudicating disputed evaluations to ensure high-consistency for training data.
Autonomous UI Review: Performed QA on novel 3D modeling pipelines where AI operated professional software. Verified actions against high-fidelity recordings.
High-Stakes Logic Validation: Conduct second and third-pass reviews on LLM reasoning, validating logic and safety alignment.

Outlier AI

LLM & Multimodal QA Analyst
APR 2023 — PRESENT Evaluation Specialist
AREAS_OF_EXPERTISE
REASONING STEM LAW PSYCHOLOGY
Complex Reasoning Evaluation: Audited model outputs for RLHF and process-supervision projects in specialized domains.
Logical Coherence Auditing: Verified step-by-step reasoning in STEM, writing, and legal contexts.

ARTIFACTS

Observability Framework

General framework for turning abstract, subjective human data into high quality rubrics with concrete observable behaviors.

VIEW DOCUMENT

Analytic Protocol

Decision making in high-risk environments using Heuer's CIA methodology. Teaching case study.

EXPLORE CASE

Alignment Hacking

Conversation logs of LLM alignment hacking through prompt injection of a fake AI data training project.

ADVERSARIAL TRAINING LOGS

EDUCATION

Library & Information Science

University of South Florida | 2023

Organizational Administration

Pasco-Hernando State College | 2021

CONTACT

dylanaoberry@gmail.com
linkedin.com/in/dylan-o-berry
Land O Lakes, FL, USA