Member of Technical Staff, Alignment

What you'll do

• Run experiments on agent behavior and improve environments to address reward hacking, cheating, unsafe tool use, and deception.
• Build and test graders, agent monitors, red-team harnesses, and safety benchmarks; secure environments against agent-caused damage.
• Conduct and publish research on oversight, control, or long-horizon autonomy, and advise partners on safety limits.

What they ask for

• In-depth knowledge of LLM safety, including attacks, red teaming, monitoring and control evaluations, and reward hacking.
• Ability to build test pipelines, learn unfamiliar codebases, and diagnose issues from logs.
• Ability to turn safety concerns into experiments and clearly interpret results.
• Experience putting research into production systems, especially post-training, distillation, or evaluation.

Nice to have

• Published work on control, dangerous capability evaluations, oversight, or interpretability.
• RLHF or RLAIF experience; public benchmark or agent task suite experience; multi-agent infrastructure experience; or AI safety/governance policy advising.

Pay

$250,000 - $450,000 USD base salary

Tools

RLHF, RLAIF

AI focus

Alignment and safety research for agentic AI, including behavior experiments, grader attacks, monitoring, red teaming, safety benchmarks, environment security, oversight, control, and long-horizon autonomy.