Soung Low

AI safety researcher and Senior Model Risk Data Scientist, with a focus on AI evaluations.

Professional Experience

• As a Senior Model Risk Data Scientist at NatWest Group, I design evaluation frameworks and lead fairness assessment standards for Gen AI and Agentic AI models, perform independent model validations, and build tooling, including a Python package for red teaming and LLM-powered governance workflows with RAG.

• Previously, as a Data Scientist at Amplifi Capital, I worked in retail lending on loan and savings product approval and pricing.

Interests

• AI safety, particularly AI evaluations. My work and training span designing evaluation frameworks for Gen AI and Agentic AI, red teaming, algorithmic bias assessment, and AI ethics and law. I am actively looking for research opportunities and collaborations, and am passionate about growing the field in Malaysia. Feel free to reach out at soung.low@outlook.com.

• Social data science, particularly NLP and quantitative textual analysis applied to political communication and representation in the Malaysian context.

Education

I hold an MSc in Applied Social Data Science with Distinction from the London School of Economics and Political Science (UK), and a BSc in Economics from Feng Chia University (Taiwan), where I ranked first in my cohort.

Training

• ML4Good (2026) — Technical AI safety bootcamp covering LLM internals, AI evals, mechanistic interpretability, and governance
• University of Edinburgh (2026) — AI and Data Ethics: ethics and law, algorithmic bias, and applied fairness tools for the financial sector
• AIS Collab (2025) — AI Alignment Track: an 8-week structured programme on the theory and practice of AI alignment
• Alan Turing Institute (2024) — Research sprint on fairness measurement in financial transaction ML models, in collaboration with Mastercard