Research Experts for Frontier AI/ML Open Questions

Mercor is working with a leading intelligence AI lab to identify the most important open questions in core AI/ML fields and to build structured knowledge bases that could meaningfully accelerate progress over the next decade. We’re looking for exceptional PhD students and PostDocs with a clear point of view on what problems truly matter in their field and the depth to define how those problems could be tackled.

You will need to fill a short form in order to be eligible for this role: You will see this in addition to the AI interview in your application process. Below is guidance for what you will need to have in order to fill the form:

leading intelligence AI lab

Eligibility Requirement

What You’ll Do

1. Identify high-impact open questions

2. Build a knowledge base for selected questions

Time commitment

Who We’re Looking For

Openings

Expected Outputs

Mercor is working with a leading intelligence AI lab to identify the most important open questions in core AI/ML fields and to build structured knowledge bases that could meaningfully accelerate progress over the next decade. We’re looking for exceptional PhD students and PostDocs with a clear point of view on what problems truly matter in their field and the depth to define how those problems could be tackled. You will need to fill a short form in order to be eligible for this role: You will see this in addition to the AI interview in your application process. Below is guidance for what you will need to have in order to fill the form: Consider the biggest open questions in your field, for example, the 10–15 questions where a breakthrough would make headline news. From this set, select those closest to your area of expertise: questions within or adjacent to your specialty, or those where you could mentor an expert toward meaningful progress. Specifically, we are looking for questions where: A major breakthrough would be widely recognized as transformative (e.g., headline news in Nature, Science, or top field-specific venues) Meaningful progress is plausible within the next decade (not purely speculative or dependent on unknown technology) The question is concrete enough that progress can be evaluated (avoid umbrella questions like “How does the ML work?”) You have the relevant expertise to assemble a comprehensive knowledge base directly relevant to the question

1. Identify high-impact open questions Propose major open questions where a breakthrough would be transformative Focus on problems that are concrete, tractable, and close to your expertise
2. Build a knowledge base for selected questions Seminal papers, key datasets, methods, recent advances, and “hidden gems” Assume an extremely strong expert all knowledge up until 6 months ago (1st October, 2025) Time commitment: ~8–16 hours per selected question PhD candidates or PostDocs from top-tier institutions Deep expertise in AI/ML/Engineering Strong judgment about significance, tractability, and research quality Ability to synthesize large bodies of literature into clear learning paths Openings: ~50 experts per domain Clearly articulated, high-impact open research questions Structured reading lists (~30–100 sources per question) Brief expert commentary on why each source and approach matters We consider all qualified applicants without regard to legally protected characteristics and provide reasonable accommodations upon request.

• A major breakthrough would be widely recognized as transformative (e.g., headline news in Nature, Science, or top field-specific venues)
• Meaningful progress is plausible within the next decade (not purely speculative or dependent on unknown technology)
• The question is concrete enough that progress can be evaluated (avoid umbrella questions like “How does the ML work?”)
• You have the relevant expertise to assemble a comprehensive knowledge base directly relevant to the question
• Propose major open questions where a breakthrough would be transformative
• Focus on problems that are concrete, tractable, and close to your expertise
• Seminal papers, key datasets, methods, recent advances, and “hidden gems”
• Assume an extremely strong expert all knowledge up until 6 months ago (1st October, 2025)
• PhD candidates or PostDocs from top-tier institutions
• Deep expertise in AI/ML/Engineering
• Strong judgment about significance, tractability, and research quality
• Ability to synthesize large bodies of literature into clear learning paths
• Openings: ~50 experts per domain
• Clearly articulated, high-impact open research questions
• Structured reading lists (~30–100 sources per question)
• Brief expert commentary on why each source and approach matters