Research Lead - Pre-training Safety

The role

The Research Lead will develop and lead work on pre-training safety to shape model capabilities and internal representations. You will build and lead a team, set research directions, and remain hands-on with code and experiments.

What you will do

• Articulate a research agenda with a clear theory of change
• Grow and lead a team of technical staff
• Lead novel research projects
• Share research findings through written content and presentations
• Mentor and coach junior team members
• Contribute to the intellectual environment and research culture
• Build a research field through grantmaking and events
• Connect research to real-world deployments

Required qualifications

• Strong existing research track record in AI or another highly technical subject (e.g. CS, math, physics)
• Deep experience with language-model pretraining, dataset construction, or controlled training experiments
• Experience building large-scale pipelines for scoring, filtering, deduplicating, and sampling training corpora
• Strong experimental judgment
• Ability to build and debug research systems directly
• A clear research agenda or a strong track record to sharpen into an agenda
• Experience leading a team, mentoring graduate students, or supporting early-career researchers
• Ability to effectively communicate novel methods and solutions to technical and non-technical audiences
• Substantive engagement with machine learning research

Nice to have

• Established publication record in AI safety
• Comfortable writing grant proposals and navigating collaborations

Benefits

• Competitive salaries
• Sizable compute budgets
• Pay for work-related travel and equipment expenses
• Catered lunch and dinner at the Berkeley office

Before you apply

• Visa sponsorship is offered for in-person employees
• Requires a CV, research direction statement, 2-3 selected works, and a note on why FAR.AI is a good home
• Process includes a 45-minute fit call, 2 technical assessments, a 3-5-day paid work trial, and reference calls

About the company

FAR.AI is a non-profit AI research institute working to ensure advanced AI is safe and beneficial for everyone. The organization conducts pre-deployment testing, independent evaluations, and publishes AI safety research.