Чётко очерченная роль в ведущей AI-компании сочетает инженерную работу над системами защиты с важной задачей безопасности; стоит учитывать дежурства и контакт с тяжёлым контентом.
Дежурства вне графика
Тяжёлый контент
Актуальные технологии
Безопасность ИИ
Влияние на систему
Мэтч & Сопровод
Для мэтча с этой вакансией нужен Plus
Подключить Plus
Описание вакансии
Текст:
Hirify AI/Оригинал
TL;DR
Protection Scientist Engineer, Integrity (AI) (SQL/Python/Machine Learning): Building production harm-detection systems, classifiers, pipelines, and interventions for AI-enabled abuse and misuse with an accent on sextortion, trafficking, national security threats, and jailbreaks. Focus on translating investigations into detection criteria, coordinating cross-functional protection strategy, and improving system effectiveness through measurement, human oversight, and language-model automation.
Location: Hybrid role based in the San Francisco office, United States; participation in an on-call rotation may require resolving urgent escalations outside normal working hours.
Salary: $220K–$425K plus equity.
Company
OpenAI is an AI research and deployment company developing general-purpose AI systems and products with a focus on safe deployment and broad human benefit.
What you will do
Define and drive protection strategy for a harm area or product and model launch, translating ambiguous risks into goals, priorities, roadmaps, and evidence-based trade-offs.
Build, evaluate, and maintain production harm-detection systems, signals, classifiers, and pipelines using SQL, Python, machine learning, language models, and shared engineering platforms.
Investigate sextortion, trafficking, national security threats, jailbreaks, and related adversarial behavior, converting case findings into detection criteria.
Lead cross-functional delivery across investigations, engineering, Policy, Operations, and Data Science, resolving dependencies and communicating recommendations, decisions, and trade-offs.
Connect detection with review, escalation, intervention, enforcement, and measurement; distinguish abusive activity from legitimate use and improve protections based on operational outcomes.
Scale reliable approaches with language models, agents, reusable tools, monitoring, documentation, and human judgement for product and model launches.
Requirements
Systems-thinking breadth across technical, analytical, investigative, and operational disciplines, with the ability to translate between specialist perspectives.
Deep expertise in at least one of sextortion, trafficking, national security, jailbreaks, or a closely related harm domain; relevant backgrounds include trust and safety investigations, counter-exploitation, threat intelligence, investigations of state-linked actors, or adversarial testing of AI safeguards.
Experience investigating adversarial behavior and translating incomplete, evolving evidence into actionable signals while accounting for uncertainty and the risk of misidentifying legitimate activity.
Proficiency in SQL and Python, including reliable and efficient production code for detection and data processing.
Ability to apply quantitative analysis and machine-learning methods to measure risks and evaluate system effectiveness, including missed harms, false positives, data limitations, and appropriate interventions.
Experience setting strategy and delivering complex cross-functional technical projects from ambiguous problem definition through completion.
Clear and concise written and verbal communication, including explaining technical findings, making actionable recommendations, and surfacing uncertainty and trade-offs.
Ability to use language models and other automation to scale analytical, investigative, or operational work with appropriate evaluation and human oversight.
Some investigations may involve sensitive content, including sexual, violent, or otherwise disturbing material.
Nice to have
Additional machine-learning or software-engineering experience.
Additional data-science or measurement background.
Culture & Benefits
Hybrid work from the San Francisco office with an on-call rotation for urgent escalations outside normal working hours.
Salary range of $220K–$425K plus equity.
Work focused on protecting users and the wider world from complex harms across OpenAI products and model launches.
Equal opportunity employment and reasonable accommodations for applicants with disabilities.
Hiring process
Applicants may be subject to background checks in accordance with applicable law.
Reasonable accommodation requests can be submitted through OpenAI’s accommodation process.
Будьте осторожны
Если работодатель просит войти в их систему, используя iCloud/Google, прислать код/пароль, запустить код/ПО, не делайте этого. Это мошенники. Обязательно жмите «Пожаловаться» или пишите в поддержку. Подробнее в гайде →