GenAI Policy Specialist - USDS

Responsibilities

About the Role
This role will sit within the USDS Policy Implementation team and focus on Generative AI (GenAI). You will champion the safe deployment of policies related to Generative AI and Machine Learning models for the US market, with a specific focus on pre-release performance evaluation, bias mitigation, and safety policy architecture. You will shape research-backed assessments to ensure our AI products-including search tools like Tako-are equitable, safe, and compliant with US-specific safety expectations before they reach our users.

Responsibilities
- Conduct the end-to-end safety and fairness evaluation of GenAI models prior to US market launch. This includes conducting AI Fairness Impact Assessments, and adhering to "go/no-go" safety thresholds.
- Partner with the Global Policy team to translate operational needs and US-specific data into requirements for new policy drafts. (Note: This role does not write the policies; it informs them).
- Act as the primary bridge between Engineering, Security, Data Science, Legal, and Product teams to translate complex technical risks into actionable policy requirements for the US market.
- Support the implementation of quantitative frameworks to assess fairness risks in LLMs and recommendation models, ensuring that marginalized and vulnerable communities within the US are protected and represented.
- Support the responsible deployment of AI-powered tools (e.g., Tako) by providing prompt-level safety guidance and optimizing product safety features to promote community well-being.