Strategic Projects Associate, Safety

Written by the T&S Network from Handshake's posting.

Runs execution of large human data programs that train and evaluate frontier AI models, focused on safety and model policy annotation. The work covers annotation frameworks, labeling workflows, and policy evaluations across a range of policy areas, used to inform model training, evaluation, and enforcement decisions.

What you'd do

• Coordinate trusted annotators on sensitive content and complex edge cases.
• Work with policy leads to bring up policy holes, unclear categories, and tricky scenarios found during labeling.
• Write and upkeep annotation playbooks, quality materials, and auditor guidance for varied datasets and legal areas.
• Analyze annotation data to find trends, systemic labeling errors, and gaps in internal safety policies.
• Coordinate with frontier AI labs and internal teams to scope projects, fix problems, and deliver policy programs.

What it asks for

• Strong academic record as a recent or soon-to-be graduate.
• Early-career exposure in trust and safety, enforcement, compliance, risk, consulting, or government is a plus (preferred).
• Sharp analytical thinking that reasons from basics, and comfort finding patterns in big data sets.
• Excellent written and verbal communication, including with senior stakeholders and customers.
• SQL proficiency (preferred).
• STEM background or exposure to data, software, or AI concepts (preferred).

Worth knowing

• Exposure to sensitive or explicit content, including violent, sexual, or otherwise disturbing material.

Our best summary of the posting, and it can be wrong. The employer's posting is what counts: read it in full on Handshakenorth_east before you apply. If we got something wrong, report this listing.