AI Researcher

We are a nonprofit raising policymaker awareness about AI risks, aiming to slow potentially dangerous research and accelerate technical safety. See the kind of work we do in our Notion or website.

As a AI Researcher Generalist at Palisade, you'll work in small teams of 1-3 people to ship projects within 2-8 weeks. You'll own projects end-to-end, from prototype to publication, collaborating with our Research Lead on direction and Engineering Lead on execution.

What you’ll do

• Pull a project from our ideas backlog or come up with your own. e.g.:
  - I think GPT-4o could do binary exploitation to the level of OSCP certificate. Concretely, I think it could hack all of crackmes.one level 3 challenges.
  - I wonder how safe the just-released o1 is 🤔
  - I wonder if https://arxiv.org/abs/2408.00761v3 really breaks the Badllama attack?
• Develop the experiment end-to-end, including its design (how to measure the thing?) and harness, then run the experiment and collect data.
• Tweet your results, and then write them up as a publication on arXiv, a landing page, or a report for policymakers.
   

Our collaboration process

• We post daily statuses for each other to keep in sync regarding our directions. Each project has a daily sync, and there’s a weekly lab demo to show off our results.
• We propose new ideas or directions by writing up a doc, sharing it, and getting comments. This enables async communication.
• Our median response time to each other is in hours, not minutes; we work in an independent and self-directed fashion. Your supervisor helps you maintain direction; colleagues help with the implementation; you keep track of your tasks and milestones.

What you’ll need

Experience level

• 4+ years of relevant technical experience, or 2+ quality research projects
• Track record of independently driving technical projects to completion
• History of mentoring others and technical leadership preferred

Core skills

• Python proficiency: it's important for coding not to get in your way while doing research
• LLM engineering: prompting, scaffolding, CoT/ToT, RAG, tool calling, agent design
• Aptitude for self-directed, high-agency work. You take initiative and contribute proactively; we don’t micromanage
• Habit of empirical over philosophical work. You go from research idea to code within hours, and from code to a tweet in 2 weeks or less
• Strong writing and presentation skills
• Advanced English language proficiency

Additional skills

This is a bucket list of relevant skills. It is not necessary to master all of them to work with us.

• LLM red-teaming: adversarial prompting, including bypassing limitations, jailbreaking, and robustness testing.
• Familiarity with AI Safety evaluations: recent work by Anthropic, Apollo, METR.
• Web development: prototype and deploy web applications and browser extensions.
• DevOps experience: reproducibility, containerization, CI/CD pipelines.
• Aptitude for cross-functional collaboration and learning. You do what it takes to ship your work.
• Motivation to conduct research that is both curiosity-driven and addresses concrete open questions in AI risk.

Hiring process

1. Apply with a CV and a cover letter
2. Optional coding test
3. Interview
4. Paid trial, 1-2 weeks

Within CV:

• Provide evidence of aptitude for self-directed high-agency work (less than 120 words)
• Provide evidence of exceptional ability (less than 120 words)
• Describe relevant experience (less than 120 words)

Compensation and benefits

• Compensation: 90th percentile market rate for your country of work authorization, capped at Netherlands market rate.
• Comprehensive global health and dental insurance
• Coworking space stipend

If you have read this text carefully, please apply through the form on our website
Also note that we can only compensate to non-Russian or Belarussian accounts