Strategy Researcher
Constellation
Date: 2 weeks ago
City: Berkeley, CA
Contract type: Contractor
About Constellation
Constellation is an independent research center that brings together people throughout the AI safety ecosystem to accelerate insight, research, and talent through better cooperation. Unlike a conference or summit, Constellation operates continuously, in the form of a physical workspace as well as conference-style talks, workshops, and training bootcamps. Continuous operation allows for relationships and conversations to develop over time, meaningfully fostering trust, collaboration, and shared insight. It also makes Constellation a natural field-building hub by rapidly inspiring, orienting, and connecting people who are newer to the field.
Based in Berkeley, CA, our shared workspace hosts over 100 people per week across dozens of AI safety organizations in nonprofits, academia, industry, and government. Hundreds of other researchers spend time at Constellation for shorter visits each year. We believe this is the strongest and highest-output network of AI safety researchers in the world; dozens of participants in past Constellation programs have gone on to safety-focused roles at companies such as METR, Redwood Research, Anthropic, Google DeepMind, OpenAI, and the US and UK Safety Institutes. Our core team currently includes :20 core members and is expected to grow to 25–30 by the end of the year. In addition, we work closely with an extended team of 5–10 collaborators.
Constellation's sole research focus is on mitigating extreme global safety and security risks from future AI systems, including human extinction. For more information, see our research focus areas and selected publications from Constellation fellows.
About The Role
Constellation is looking for people to do semi-independent research on AI safety macrostrategy, part-time or full-time, for 1-3 months.
The goal is to produce a broad overview of the most promising interventions for mitigating catastrophic risks from the development of transformative AI. We expect a focus on analyzing, aggregating, and synthesizing existing proposals, as opposed to developing novel interventions. We are most interested in work describing portfolios of interventions that might collectively mitigate all such risks, but would consider proposals from candidates interested in focusing on a broad subset of TAI risks (e.g. “loss of control” or “concentration of power”) or interventions (e.g. policy and/or governance interventions).
The output of the engagement will be a formal or informal publication (e.g. blog post) which documents your work. Your analysis will inform Constellation’s strategic planning, including which field-building programs we prioritize.
Work will be fairly independent. We will supply a more specific prompt in line with the goals described above, and offer periodic feedback and questions, but you will be expected and encouraged to pursue your work as you see fit within the broad constraints of the prompt.
Skills
You’re ideal for this engagement if you have:
We expect to structure this engagement as an hourly contract. Rates will depend on experience, but will likely be between $50/hr and $150/hr.
This is a temporary position for 1-3 months with a flexible schedule. Full-time or part-time options are available. We are primarily an onsite team based in Berkeley, and the Constellation workspace hosts many top safety researchers. We will, however, consider remote candidates for this particular engagement.
If there is a strong mutual fit, we may be interested in extending the engagement or converting this to an ongoing, full-time role, though this scenario should be considered exploratory and is also dependent on other factors. We may be able to sponsor visas for such cases. Ongoing, full-time roles would be onsite in Berkeley, CA. Our office is located under 20 seconds from the nearest BART (metro) stop, and on-site parking is also available.
We value diversity in all respects and base our hiring decisions on the needs of the organization and individual qualifications. We welcome applicants from all backgrounds, regardless of race, color, religion, sex, sexual orientation, gender identity or expression, national origin, age or disability.
Constellation is an independent research center that brings together people throughout the AI safety ecosystem to accelerate insight, research, and talent through better cooperation. Unlike a conference or summit, Constellation operates continuously, in the form of a physical workspace as well as conference-style talks, workshops, and training bootcamps. Continuous operation allows for relationships and conversations to develop over time, meaningfully fostering trust, collaboration, and shared insight. It also makes Constellation a natural field-building hub by rapidly inspiring, orienting, and connecting people who are newer to the field.
Based in Berkeley, CA, our shared workspace hosts over 100 people per week across dozens of AI safety organizations in nonprofits, academia, industry, and government. Hundreds of other researchers spend time at Constellation for shorter visits each year. We believe this is the strongest and highest-output network of AI safety researchers in the world; dozens of participants in past Constellation programs have gone on to safety-focused roles at companies such as METR, Redwood Research, Anthropic, Google DeepMind, OpenAI, and the US and UK Safety Institutes. Our core team currently includes :20 core members and is expected to grow to 25–30 by the end of the year. In addition, we work closely with an extended team of 5–10 collaborators.
Constellation's sole research focus is on mitigating extreme global safety and security risks from future AI systems, including human extinction. For more information, see our research focus areas and selected publications from Constellation fellows.
About The Role
Constellation is looking for people to do semi-independent research on AI safety macrostrategy, part-time or full-time, for 1-3 months.
The goal is to produce a broad overview of the most promising interventions for mitigating catastrophic risks from the development of transformative AI. We expect a focus on analyzing, aggregating, and synthesizing existing proposals, as opposed to developing novel interventions. We are most interested in work describing portfolios of interventions that might collectively mitigate all such risks, but would consider proposals from candidates interested in focusing on a broad subset of TAI risks (e.g. “loss of control” or “concentration of power”) or interventions (e.g. policy and/or governance interventions).
The output of the engagement will be a formal or informal publication (e.g. blog post) which documents your work. Your analysis will inform Constellation’s strategic planning, including which field-building programs we prioritize.
Work will be fairly independent. We will supply a more specific prompt in line with the goals described above, and offer periodic feedback and questions, but you will be expected and encouraged to pursue your work as you see fit within the broad constraints of the prompt.
Skills
You’re ideal for this engagement if you have:
- Extensive domain expertise in AI safety, especially broad expertise thinking holistically about threat models and mitigations
- Exceptional epistemic clarity, including reasoning with and communicating about uncertainty
- The ability to identify and build from the fundamental assumptions or elements of a problem
- Experience reasoning about practical constraints such as decision-making processes, organizational dynamics, international politics, and project planning
- The ability to create clarity for others in complex conceptual terrain (through precise language, intuitive models and frameworks, etc.)
- Very clear and concise writing
- Experience successfully managing your own output on long projects with limited direction
We expect to structure this engagement as an hourly contract. Rates will depend on experience, but will likely be between $50/hr and $150/hr.
This is a temporary position for 1-3 months with a flexible schedule. Full-time or part-time options are available. We are primarily an onsite team based in Berkeley, and the Constellation workspace hosts many top safety researchers. We will, however, consider remote candidates for this particular engagement.
If there is a strong mutual fit, we may be interested in extending the engagement or converting this to an ongoing, full-time role, though this scenario should be considered exploratory and is also dependent on other factors. We may be able to sponsor visas for such cases. Ongoing, full-time roles would be onsite in Berkeley, CA. Our office is located under 20 seconds from the nearest BART (metro) stop, and on-site parking is also available.
We value diversity in all respects and base our hiring decisions on the needs of the organization and individual qualifications. We welcome applicants from all backgrounds, regardless of race, color, religion, sex, sexual orientation, gender identity or expression, national origin, age or disability.
How to apply
To apply for this job you need to authorize on our website. If you don't have an account yet, please register.
Post a resume