Research Operations, Reinforcement Learning at Anthropic in San Francisco, CA

Job description

• Frontier research runs on more than research alone. Behind every model launch, review, and cross-team program is a web of decisions, priorities, and follow-through that keeps hundreds of researchers moving in the same direction. Research Operations owns that work: we bring clarity to ambiguity, provide stability during change, and continually deepen our domain knowledge so that researchers can spend their time on the problems only they can solve. It is one of the highest-leverage functions at Anthropic.
• We are hiring an experienced operator to embed with the leadership of our reinforcement learning (RL) organization, the team responsible for training Claude to be more capable, more reliable, and more aligned. You will shape what leadership spends its time on, make sure priority decisions reach the right decision-makers and get made quickly, and clear the path so leaders can focus on the work that matters most. You will report to the Manager of Research Operations and work daily with RL org leaders and their teams.

Responsibilities

• Own leadership's time. Shape the week around the leaders' specific, most timely priorities, set agendas so meetings are spent on decisions that move the org forward, and cut what isn't critical so leaders stay focused on the areas that are highest-leverage for them.
• Triage what reaches leadership. Resolve what you can, route what belongs elsewhere, and thoughtfully escalate the rest.
• Enable wise decision-making and follow through. Give leaders what they need to decide quickly, communicate those decisions to collaborators, and track commitments through to completion.
• See problems coming. Surface resourcing gaps, technical bottlenecks, and stalled action-items before they become blockers.
• Scale leadership voice. Meetings are ephemeral; docs are forever. Draft materials for org-wide announcements, programming, and planning sessions. Protect leadership’s time for creating durable comms that set direction in turbulent times.
• Make the org run better. Find the friction that keeps recurring and find the people who can fix the process behind it.
• Be a thought partner. Offer honest, candid counsel to leaders, including crisp and kind disagreement.

Requirements

• Care deeply about Anthropic's mission and let it underpin all judgment calls you make across the org.
• Technical fluency to engage substantively with researchers: you can follow a discussion about an RL training run, read a set of eval results and form reasonable hypotheses about what they mean, and hold your own in a technical conversation. You don't need to have trained a model yourself, but you should understand the model training process and have the depth to work closely with leading researchers.
• Clear writing: you turn dense, jargon-heavy material into summaries and plans that technical collaborators across the RL org and beyond can act on
• An eye for great dashboards: you can look at data visualizations and scope improvements that make lessons at-a-glance clearer and grounded.
• Strong follow-through and attention to detail; nothing falls through the cracks on your watch.
• Ability to push back on senior stakeholders to offer strategic guidance as needed, disagree without causing discord, and influence without direct authority.
• Comfort with ambiguity and a fast-changing environment; you stay steady when the pace is relentless and the stakes are as high as they can possibly be.
• Experience in an operations, chief of staff, program management, or comparable role supporting senior leaders in a fast-moving technical organization
• High integrity. This role sits close to decisions that execute on Anthropic’s mission, and you care about getting them right.
• Experience supporting a senior leader directly in a chief of staff or related function
• Time in a research institution, lab, or other research-heavy organization
• Track record of anticipating what colleagues and leaders will need and acting on it before being asked
• Comfort in an environment where the right answer changes hour to hour
• Experience moving routine operational work to AI agents, or the ability to script and build small tools
• Familiarity with AI safety or other existential risks.

Skills

Required

• Clear writing
• An eye for great dashboards
• High integrity

Benefits

• The annual compensation range for this role is listed below.
• For sales roles, the range provided is the role’s On Target Earnings ("OTE") range, meaning that the range includes both the sales commissions/sales bonuses target and annual base salary for the role.
• $210,000 - $260,000 USD