Atlas Computing is incredibly excited to partner with Coefficient Giving to host Founders-in-Residence, an evolution of what was previously the Navigators Incubator. We’re already collaborating on the full program design, recruiting, and selection, and look forward to hosting and supporting the fellows once the program begins. Read more about the program below, or go straight to our applications.
What the Founder-in-Residence program is
We know there are people with deeply relevant domain expertise who want to mitigate the most acute, catastrophic risks to society posed by growing AI capabilities.
Our job is to lower the barriers to entry. That means both reducing the cost of leaving your current career path and the risk that a good idea never gets off the ground. We’re also trying to raise the odds that the effort you launch succeeds. Each founder joins with a relevant risk they care about, and we help them develop and refine a rough direction into a de-risked, funding-ready concrete intervention. This involves pressure-testing it against the experts, building customer or user relationships, and assembling a team to execute it. At the end of the program, founders will have put together an excellent proposal, identified a founding team, and we’ll get them pitches with funders like Coefficient Giving for initial support and a clear path to scaling up the project with grants of $10 million or more.
The program runs about six months, but we hold that loosely — a founder is done when their effort is ready to stand on its own, not when a calendar says so.
Cohort 1 scope: AI Infrastructure Security
We want founders to apply because they care about a risk, not a solution. So naming which risks are in scope is critical.
The first 10 Founders-in-Residence we’re hiring will be focused on any of six very different aspects of securing AI infrastructure:
Securing AI from exfiltration: Protecting model weights and algorithms from theft.
Securing AI integrity: Protecting models from tampering via data poisoning or other attacks.
Disconnected model safety: Preventing models that are not being served by AI labs from being used for catastrophic purposes.
Secure compute verification: Verifying that large-scale compute workloads are compliant with law or international agreements.
Protocols for using untrusted AI: Also called AI control — how to use untrusted AI to do useful work.
Detecting and responding to rogue deployments: What to do if we fail to contain transformative AI and it gets deployed (or deploys itself) without oversight.
You can read more about these problems here. We believe these problems are important and urgent, as transformative AI clearly needs to be secured against threat actors like nation-states, terrorists, insider threats, and even misaligned AIs themselves. These are threat actors who may have enormous resources and capabilities, intentions to harm large numbers of people, or significant legitimate access. We expect these problems in the next few years will be the hardest, highest-stakes security problems anyone will work on in their lifetimes.
What success looks like
We measure the program by where Founders go next. Over time, we hope they become:
- Founders and DRIs who lead object-level interventions — the people directly building the organizations, tools, and efforts that reduce catastrophic risk from AI. We don’t care if the new entity is a for-profit or non-profit, as long as it’s the best structure to mitigate the risk.
- Grantmakers and grant managers who can strategically direct serious capital or manage teams of funders across this field.
- Advisors, mentors, and recruiters for the next wave of founders working on these problems — multiplying their impact by helping others launch and aim.
We’re not looking to maximize follow-on funding or be a resume filler. We’re looking for successful candidates to be people the field builds around.
And since we don’t want to incentivize ourselves into pushing people toward suboptimal solutions, we’re not taking equity in any of the startups.
Two tracks: Founders-in-Residence and Field Strategists
Alongside the incubator, Atlas is also hiring Field Strategists — and we’re running both searches at once, with open applications for each.
The two roles are complementary, and the difference is a bit like choosing an undergraduate major. You apply to one or the other, and the applications are different, but you’re not locked in.
Founders focus on burning through solution directions until they find one that’s viable and helpful.
The interview process for founders focuses on your ability to generate and refine a proposal for an intervention, but you will not be required or even expected to focus on that solution. (See the Founder-in-Residence application.)
The work is about building — taking a single object-level project and iterating a rough idea into a fundable, executable plan. Founders will learn and be coached on field strategy for their problem of choice, develop a proposal in partnership with Atlas leadership and funders, and transition out of Atlas as soon as their project is ready to scale.
A founder is more likely to find it more efficient and faster to focus on one intervention, and would prefer to own execution on the idea themselves to make sure it’s done to their standard.
Field Strategists focus on mapping a field’s incentives, finding where the leverage is, and de-risking high-impact interventions.
The interview process includes a speed-run of our field strategy process, but also covers how you think and how you work. (See the Field Strategist application.)
The field strategist holds an ongoing position at Atlas, building the skills and praxis for this approach for multiple efforts simultaneously and in series.
A field strategist is more likely to find it more interesting and energizing to pursue multiple interventions at once, and find handing off interventions to someone else for execution to be a relief.
Note on Field Strategy topics: We’re also hiring Field Strategists to work on critical infrastructure cybersecurity and epistemic security, in addition to the six problems above in AI infrastructure security.
Regardless of track, everyone is pulled into the shared program curriculum including guest sessions, context building, skill-set training, idea red-teaming, and setting+achieving milestones.
Please send questions to join@atlasignota.org — we’ll try to answer questions briefly there, and will post a follow-up with FAQs if/when we observe any patterns emerging.
Who we’re looking for
Everyone’s looking for exceptional individuals with great communication skills who can handle ambiguity... blah blah.
We value transparency: here’s most of our initial screening criteria:
Founder Baseline: We’re looking for individuals with one of the following backgrounds (or similar):
Experience working on cybersecurity in a national security context (especially building SL5 facilities, intelligence, and offensive security).
Experience working on ML security research (tamper resistance, backdoors, password locking).
Experience working in a relevant industry context — e.g. at a frontier AI lab, or security at a hyperscaler, hardware company, or government contractor.
Exceptional capability in a relevant area — e.g. academic security research, security entrepreneurship, developing security standards, or AI hardware or firmware.
Field Strategy Baseline: We’re looking for individuals who
have an unreasonably large, relevant professional network and/or have an unreasonable understanding of how decisions are actually made in some relevant industries (policymaking, regulation, startups, F50 companies, frontier AI labs, etc).
are great at getting into technical details, getting conversations with incredibly busy people, and getting people to tell you things they wouldn’t put in writing.
Bonus: We’re especially excited for candidates who:
have a habit of doing non-technical coordination work despite a technical background.
have a deep conviction that the path to better outcomes runs through changing the minds of decision-makers.
understand what it would mean to be responsible for mitigating a particular risk from AI, and you want it anyway (because right now, no one is).
Lastly, Atlas is a good fit for you if:
you’re not in a place in your life to leave your career to join a normal incubator or fellowship program with a small stipend, but you feel like you have to do something about AI risk, and you’re not yet sure what.
Anti-criteria: You’re probably not a good fit if:
- you already have the answer to the problem you want to solve, and you want funding more than a thinking partner
- you want a stable, well-scoped role with a clear roadmap handed to you.
- you’re drawn to AI risk in the abstract but don’t want to own a specific, concrete problem.
- you don’t (yet) have a broad professional network or deep domain expertise. We’ll take young people who score well on these relative to their career stage, but won’t be doing things like running an internship program).
Compensation
Founders-in-Residence and Field Strategists will be full members of the Atlas team with salary and generous benefits. Compensation is designed to match private-sector pay, up to a generous cap, so that taking this on won’t mean a pay cut for the people best positioned to do it.
Apply
Applications are open now — see the Founder in Residence application and the Field Strategist application. We’re keeping both open on a rolling basis with no application deadline for the foreseeable future, though we expect to pause accepting new candidates once we’ve filled the current cohort, so it’s worth applying sooner rather than later.
And if you know someone who’d be a strong fit, we’d be grateful if you’d share this with them.