Research Scientist – Science of Evaluation

NATIONAL(4 days ago)

HybridFull TimeSenior$82,550 - $184,150Science of Evaluation

About this role

A Research Scientist role at the AI Security Institute’s Science of Evaluation team focused on improving measurement and forecasting of frontier AI capabilities, particularly for long-horizon agents and large inference budgets. The position supports government-facing research to inform policy and works closely with internal teams and external partners. Researchers are given substantial autonomy to shape independent agendas within a mission-driven public-interest organisation.

View Original Listing

Required Skills

Evaluation Methods
Long-Horizon
Experiment Design
LLMs
Agents
Forecasting
Transcript Analysis
Data Engineering
Statistical Analysis

Qualifications

PhD in Technical Field

About NATIONAL

gov.uk

GOV.UK is the UK government’s official online portal that centralises public services and authoritative information, from taxes, benefits and passports to driving, visas and legislation. Maintained by the Government Digital Service, it brings content from departments and agencies into a single, accessible site with transactional services, forms and practical guidance for citizens, businesses and professionals. The site emphasises clarity, accessibility and security, offering search, sign-in tools and up‑to‑date policy and regulatory information to help users complete tasks. It also publishes open data, guidance for developers and follows user‑centred design and publishing standards to continually improve government services.