Head of Technical Standards and Best Practice

Primary objective: Lead AISI's work to define the technical best practice needed to manage the most serious risks from frontier AI development, and to get it adopted. 

You'll work with AISI leadership to set the direction for this work and be accountable for delivering it. You'll lead a diverse team of technical experts and policy specialists who turn AISI's research into best practice, and take it into the forums where it can shape what frontier AI developers do. 

Core responsibilities:

• Set direction and strategy (20%)
• • Decide which areas of best practice AISI should work on, and in what order, based on risk, AISI's expertise and appetite among developers.
  • Own AISI's long-term strategy in this area and make the case inside and outside AISI for why it matters.
• Lead best-practice content (40%)
• • Oversee researchers and experts across AISI as they develop best practice, and bring their expertise together into clear, usable documents.
  • Make sure each document sets out specific, actionable practices: what developers should do, and how anyone could check it has been done.
• Technical engagement and adoption (25%)
  • Lead technical engagement with frontier AI developers, evaluators, standards bodies and government partners to test and improve drafts.
  • Oversee publications and company engagement, choosing the best route to adoption for each area.
• Build and lead the team (15%)
  • Line manage and grow the core team and set up a repeatable process from first draft to publication.

Example projects:

• Take AISI's best practice on monitoring and oversight from draft through testing with developers to publication.

• Lead best practice on which metrics developers should report as AI starts to speed up AI research, and what should be shared publicly or confidentially.

• Choose the next priority areas and set up the teams to deliver them.

Reporting and growth:

You'll report to a Deputy Director and line manage the core team. We're looking for a long-term lead: as the team grows, so will the role’s remit. 

What we are looking for

We're flexible on the exact profile and expect successful candidates will meet many, but not necessarily all, of the criteria below. Depending on experience, we will consider candidates at Level 5 or Level 6, and exceptional candidates at Level 4.

Essential:

• Strong technical understanding of frontier AI systems and of the practices used to make them safe and secure, gained for example through 5+ years in AI security research or engineering, or AI evaluation.

• A track record of thought leadership on safer frontier AI development practices. For example: developed an industry standard, a published safety framework, or internal policies that engineering teams followed.

• Experience leading a team, or managing a large programme of work involving several organisations.

• Experience building credibility with senior technical people within and outside your own organisation.

• Clear, precise writing for both technical experts and senior decision-makers.

Desirable:

• Experience working at, or closely with, a frontier AI developer.

• Experience writing or reviewing safety cases, either for AI or in a safety-critical industry such as nuclear, aviation or medical devices.

• Expertise in one of our priority areas: monitorability, documenting and reporting risk, or transparency about AI research automation.

• Experience with formal standards development (for example ISO or NIST processes) or industry consortia.

Example backgrounds:

• A senior researcher or engineer from a frontier AI company's safety, alignment or security team who wrote internal safety policies or frameworks.

• A researcher at a third-party evaluation or AI safety organisation who has led published frameworks for evaluating or governing frontier AI.

• A safety or security engineer from a safety-critical industry who has led safety case or standards work and has moved into AI.

• A lead at a standards body or industry consortium who has been hands-on drafting technical content specifically on frontier AI development, not only running the process.

Core requirements

• You should be able to spend at least 4 days per week working with us.

• You should be able to work from our London office roughly 3 days per week.

• No security clearance is needed to apply. Eligibility for Security Check (SC) clearance is preferred, but not essential. More detail on clearance eligibility can be found on the UK Government website: National security vetting: clearance levels - GOV.UK.
  

Interview process

Candidates should expect to go through some or all of the following stages once an application has been submitted:  

• Initial interview  
• Take home test 
• Second interview and review of take home test 
• Third interview 
• Final interview with members of the senior team