Research Scientist, Applied White-Box Methods

FAR.AI

. Take ownership of and accelerate the Applied White-Box Methods team's research agenda .

Posted 9/23/2026full-timeBerkeley • California • United StatesMid-LevelSenior💰 $150,000 - $250,000 per yearWebsite

Core Competencies

Role fit

Core Competencies

Use this summary to align your resume positioning with the role.

Demonstrates expertise in AI safety and white-box methods, with a strong ability to develop and evaluate innovative techniques for improving AI systems. Engages effectively with both technical and non-technical audiences while contributing to the AI alignment community.

Highest-signal resume keywords

White-Box Method ApplicationAI Safety ResearchEvaluations and AI ControlPhD in Computer Science or Related FieldApplied Machine Learning Experience

ATS Keywords

Tailor your resume

Applicant Tracking System Keywords

Tip: use these terms in your resume and cover letter to boost ATS matches.

Hard Skills

Activation ExplainersSteeringAttributionInfluence FunctionsReinforcement LearningPost-Training of LLMsGoodhart-Resistant MetricsLong-Context Agentic CodingResearch TamperingReward Hacking

Soft Skills

Clear CommunicationCollaboration

Industry Keywords

AI AlignmentRed-TeamingFrontier-Scale DeploymentLarge-Scale ModelsTechnical Writing

About the role

Key responsibilities & impact

Take ownership of and accelerate the Applied White-Box Methods team's research agenda
Develop, evaluate, and demonstrate methods leveraging model internals to improve AI safety
Stress-test white-box methods on real-world tasks, including long-context agentic coding
Study applications including white-box control, evaluation awareness, training-dynamics shaping, alignment, reward hacking, sandbagging, and research tampering
Define realistic evaluations with Goodhart-resistant metrics
Develop methods using realistic, large-scale models and evaluations
Compare white-box methods against black-box methods and activation probes
Ensure practical monitors and interventions are simple and efficient enough for frontier-scale deployment
Publish findings broadly and engage with the AI alignment community
Propose new research directions within the team's agenda
Attend relevant conferences and community events
Collaborate with national AI safety institutes, frontier model developers, and top academics
Complete interviews with technical staff followed by a paid work trial of up to one week

Requirements

What you’ll need

Hands-on experience applying at least one white-box method to a real model, such as activation explainers, SAEs, steering, attribution, influence functions, or probes, and an informed view of its limitations
Track record in AI safety through a paper, fellowship project, or substantive public writing
Experience with evaluations, AI control, red-teaming, reinforcement learning, or post-training of LLMs
Ability to communicate novel methods and results clearly to technical and non-technical audiences
PhD or several years of research experience in computer science, machine learning, physics, statistics, or a related field
Previous experience in applied ML for fields such as biology, chemistry, materials science, or robotics
For researchers new to LLM research: ability to explain how previous research experience could be leveraged for this work and how they are engaging with technical AI safety research
Willingness to work in person in Berkeley or relocate for the in-person role, unless considered for an exceptional remote arrangement

Benefits

Comp & perks

Health insurance: 94% of insurance premium paid by Organization commencing within 1 month after your start date
401(k) plan with up to 2% match
25 days Paid Time Off per year, accrued weekly
Up to 10 days of paid sick leave per year
Paid Bereavement, Family, Medical and Pregnancy Disability Leave
Work computer and WFH stipend provided for eligible employees
Catered lunches and dinners on workdays at the Berkeley office
Work-related travel and equipment expenses paid
Visa sponsorship for in-person employees