Applied Research - Computer Use

Skills

About the Role

You will build post-training and evaluation systems for multimodal agents operating across browsers, desktops, and applications. You will turn workflows into reliable training environments, develop safe sandboxed agents, and evaluate recovery, completion, permissions, and generalization.

Requirements

• Experience with multimodal agents, computer use, browser or desktop automation, RL, imitation learning, HCI, embodied agents, or long-horizon agent evaluation.
• Experience building and debugging model-training, post-training, agent, evaluation, or adjacent experimental loops.
• Strong experimental judgment, including meaningful baselines, trace inspection, and failure attribution.
• Strong Python skills.
• Ability to build artifacts that other researchers and engineers can run, inspect, and extend.
• Ability to own ambiguous problems and keep partners and teammates informed.

Responsibilities

• Build long-horizon computer-use environments from real workflows.
• Train agents to recover from interface changes, partial failures, and ambiguous states.
• Develop evaluations for completion, correctness, efficiency, permissions, and escalation.
• Build multimodal environments using workflows, demonstrations, tools, interface state, and outcome criteria.
• Develop browser, desktop, and cross-application agents with sandbox and boundary enforcement.
• Train and evaluate agents using demonstrations, interventions, production failures, synthetic tasks, and verifiable outcomes.
• Run experiments on recovery and generalization across applications, layouts, permissions, and deployment contexts.

Benefits

• Flexible work in San Francisco, New York, London, or remote
• Visa sponsorship
• Relocation support
• Professional development budget
• Team off-sites
• Conference attendance