Technical Writer, Frontier AI Best Practice
Мэтч & Сопровод
Для мэтча с этой вакансией нужен Plus
Описание вакансии
Location: London, UK; hybrid working with options for similar government offices in Birmingham, Cardiff, Darlington, Edinburgh, Salford or Bristol. At least 4 days per week must be spent working with AISI; occasional remote work abroad is available.
Annual salary: £65,000–£145,000, including base salary and technical allowance, with an additional 28.97% employer pension contribution on base salary.
Company
The is a UK government research institute focused on understanding advanced AI risks and translating research into technical best practice, standards and policy action.
What you will do
- Work with researchers and the Technical Standards and Best Practice team to define and structure best-practice documents.
- Own delivery of documents by coordinating technical input, editing content, and ensuring that practices are coherent, specific, testable and technically accurate.
- Explain the technical rationale for practices using research findings, threat models and evaluation results.
- Develop consistent style, templates and terminology, and manage internal and external expert review rounds.
- Write public explainers and technical blog posts for developers, policymakers and the public.
- Manage multi-stakeholder drafting and adapt technical material for both research and non-technical policy audiences.
Requirements
- Ability to understand nuanced technical AI safety work and explain its claims and limitations accurately.
- Ability to write precise, structured documents explaining what organisations should do and why.
- Ability to explain technical topics clearly and engagingly to non-technical readers.
- Relevant experience may include writing or co-writing AI research papers, technical communications with researchers, technical guidance or standards, complex-field documentation, or public and policy writing on technical topics; candidates do not need all examples.
- Ability to work with AISI at least 4 days per week.
- Successful candidates must complete a criminal record check and obtain BPSS clearance before appointment.
- Eligibility for SC clearance is preferred but not essential; CTC eligibility is also strongly preferred.
Nice to have
- Familiarity with monitorability, safety cases and risk reports, or AI research automation.
- Extensive experience communicating AI security concepts to policy audiences.
- Experience leading the writing of multi-author documents, including gathering expert input, resolving conflicting feedback and taking documents to publication.
- Familiarity with the work of frontier AI developers, evaluators or standards bodies.
Culture & Benefits
- Work with researchers, engineers, communications, international and strategy specialists, government experts and leading AI companies.
- Access to pre-release frontier models, substantial compute and operational research support.
- Opportunities to influence frontier AI governance, publish externally and collaborate with international and third-party research organisations.
- Five days off, annual learning and development stipends, conference and external collaboration funding, and at least 25 days of annual leave plus public holidays and team breaks.
- Hybrid work, occasional remote work abroad, work-from-home equipment support, paid parental leave and pension contributions.
- The role normally has no line-management responsibility, although responsibility may grow with the team.
Hiring process
- Depending on the application, candidates may complete an initial interview, technical take-home test, second interview with test review, third interview and final interview with senior team members.
- Application examples and statements must be truthful, factually accurate and based on the applicant's own experience; plagiarism or presenting AI-generated work as personal experience may lead to withdrawal.
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 →