I'm interested in building better ways of evaluating the safety of frontier AI models, and conducting research to understand their impact on society. Developing AI safety evaluation methodologies, conducting ML research, and building scalable systems to facilitate and productionise research.
EXPERIENCE
Senior Data Scientist
The Alan Turing Institute
2022 - Present
Data Scientist 2021-2022
Working with AI researchers, social scientists, and policy experts on AI Safety/Online Safety research.
Building methods and infrastructure for sociotechnical AI Safety evaluations.
Safety testing pre-release frontier models.
Conducting & applying technical research, fine-tuning and evaluating LLMs.
Built large-scale data/ML pipelines for analysing online abuse.
Data Scientist
Curve Analytics
2021
Developed and deployed search traffic ranking solutions and recommender systems at data science consultancy.
SKILLS
Python
PyTorch
HF Transformers
Sklearn
Flask
SQL
Azure
AWS
Docker
Git
Machine Learning
Deep Learning
Transformers
NLP
Tech Policy
EDUCATION
MEng (1st)
Computer Science with Innovation
University of Bristol
2016 - 2020
Units taken include:
Applied Data Science
Applied Deep Learning
Machine Learning
Software Product Engineering
Design & Systems Thinking for Innovation
🎭 LLM Disinformation
Evaluating LLMs for malicious info ops
Measuring compliance and "humanness" of AI-generated election disinformation.
🐙 Prompto
Facilitating LLM experiments
Python library for asynchronous querying of LLM endpoints.
🦤 DoDo Learning
Cross-domain abuse classification
Exploring domain-demographic transfer performance in language models for abuse classification.
🔭 Online Harms Observatory
Tracking online abuse
Building pipelines and deploying language models to deliver real-time analytics on online harm.