Kate (catherine) M.

Key Skills

Software

Internal/Proprietary Tooling

CVAT

Label Studio

Top Subject Matter

Affective Intelligence

AI Safety and Alignment

Behavioural Psychology

Top Data Types

Text

Document

Top Task Types

RLHF

Text Generation

Red Teaming

Evaluation/Rating

Question Answering

Prompt + Response Writing (SFT)

Classification

Freelancer Overview

First Class Masters in Music Therapy (clinical psychology specialism, University of Limerick), integrating developmental psychology, psychodynamic and psychosocial theory and evidence-based research. Currently working as AI Data Operations & Quality Lead at Emotionise AI, applying psychological theory to affective alignment, evaluation rubric design, and adversarial stress-testing for SLM development. Led ground-truth remediation recovering SLM accuracy from 53% to 85%.

Labeling Experience

AI Data Operations & Quality Lead, Emotionise AI

TextRLHF

As AI Data Operations & Quality Lead at Emotionise AI, I designed and implemented HITL pipelines and Affective Labelling frameworks for SLM development, using Excel and Power BI for quality monitoring and accuracy tracking. Core work spanned ground-truth remediation, adversarial red teaming, evaluation rubric design and affective taxonomy development for RLHF and model alignment. • Designed and deployed the foundational Affective Labelling Taxonomy for SLM emotion recognition across nuanced empathy registers - the primary training framework for the model’s emotional intelligence layer. • Led ground-truth remediation diagnosing synthetic data decay and rebuilding the quality pipeline - recovering SLM accuracy from 53% to 85%. • Produced structured adversarial stress-testing scripts, hallucination triage protocols, and multi-axis evaluation rubrics for RLHF integration. • Applied EU AI Act Article 14 (Human Oversight) requirements directly to data labelling pipeline design and evaluation protocols.

2024 - Present

Education

Work History