Protection Scientist, Engineer

This is a fully remote position, open to applicants in United Kingdom.

📋 Description

• Identify and analyze the misuse of the client's products, focusing on emerging types of abuse.

• Create and develop systems that proactively detect and address abuse in AI products.

• Implement comprehensive abuse monitoring for newly launched products while maintaining oversight for existing ones.

• Prototype and develop defensive systems to counter high-risk harms.

• Address and investigate critical escalations that are not captured by current safety protocols.

• Gain a thorough understanding of products and the data associated with them.

• Define and execute abuse monitoring criteria for new product launches.

• Enhance monitoring operations and automate subtasks for ongoing products.

• Prototype and refine production systems for abuse detection, review, and enforcement.

• Collaborate with teams across Product, Policy, Operations, Investigative, and Engineering to assess risks, formulate responses, and develop scalable tools.

⛳️ Requirements

• Must be able to work remotely on GMT and be located in the UK.

• Background in quantitative analysis and coding, including proficiency in statistics and metrics.

• Expertise in Python and SQL.

• Experience in identifying and, preferably, enforcing actions against bad actors using scalable tools.

• Willingness to participate in an on-call rotation approximately once every quarter, addressing urgent escalations outside of regular working hours.

• Ability to quickly switch contexts across various domains, modalities, and abuse areas, including violent activities and child safety.

• Background in machine learning and classification, particularly concerning novel or poorly understood behaviors (desired but not mandatory).

• Experience in scaling and automating processes, particularly with language models (desired but not mandatory).

• Familiarity with working alongside scaled labellers or reviewers (desired but not mandatory).

• Knowledge of AI safety, child safety, abuse by state actors and other malicious entities, digital mental health, fraud abuse, radicalization/persuasion/grooming, or extremist activities and groups (desired but not mandatory).

🏝️ Benefits

• Full-Time (40 hours per week) contract position.

• Remote role based in London, UK.

• Participation in an on-call rotation approximately once per quarter, including urgent escalations during evenings and weekends.