Research Manager, AI Safety

About the role

You will own the end to end management of high impact research projects inside the Cambridge Boston Alignment Initiative, translating strategic research goals into concrete execution plans for technical fellows. You will act as a primary liaison between renowned external mentors and our internal fellows, synthesizing feedback and ensuring research activities remain coherent and rigorous. A core part of your role involves designing and running structured support mechanisms such as developmental coaching and methodological workshops to elevate the quality of ongoing investigations. You will proactively identify and mitigate risks to research progress by anticipating obstacles in experiments, analysis plans, and theoretical assumptions. Additionally, you will steward the creation of public facing research artifacts that clarify our technical approaches and governance insights for broader audiences. You will also safeguard the integrity of our evaluation frameworks, ensuring that results are reproducible and that findings withstand external scrutiny. Through these duties, you will directly shape the trajectory of impactful AI safety research and the careers of emerging researchers.

Key facts

What you'll do

Coordinate complex research workflows by running detailed project plans that track milestones, dependencies, and resource allocation for multiple concurrent fellow projects.

Conduct rigorous 1-1 coaching sessions where you analyze research outputs, diagnose conceptual gaps, and guide fellows through iterative refinement of their hypotheses.

Construct and maintain structured scaffolds such as literature reviews, assumption trees, and causal models to support fellows in organizing large bodies of technical information.

Facilitate deep technical dialogues between fellows and their assigned mentors, ensuring alignment on research objectives, success criteria, and risk management strategies.

Implement robust evaluation frameworks that combine qualitative assessments with quantitative metrics to measure research progress and outcome quality.

Curate dynamic resource libraries tailored to fellow expertise levels, including papers, tooling recommendations, and replication studies relevant to AI safety subdomains.

Lead targeted speaker events and reading group sessions that connect fellows with leading practitioners and emerging findings in alignment, interpretability, and governance.

Integrate feedback from mentor communications into updated research briefs that clarify priorities and adjust experimental designs in response to new information.

Oversee the candidate selection pipeline by developing interview protocols, scoring rubrics, and calibration sessions to build cohorts with strong epistemic discipline.

Produce concise weekly synthesis documents that translate dense technical literature into accessible insights for both technical and non technical stakeholders.

Monitor external developments in AI control, formal verification, and policy landscapes to identify opportunities for collaboration or new research directions.

Support special projects such as evaluation design, tool development, and documentation systems that enhance the scalability and reproducibility of fellowship outputs.

Champion a culture of intellectual humility and rigor by modeling best practices in research critique, transparency, and preregistration where applicable.

Collaborate with operations and leadership teams to refine program infrastructure, ensuring that fellowship cycles scale smoothly without degradation in mentorship quality.

Requirements

You hold a Bachelor's or Master's degree in a technical field with substantial exposure to advanced AI systems, machine learning, or computational theory.

You have direct experience conducting or managing research that involves quantitative analysis, experimental design, or formal methods relevant to AI systems.

You are deeply familiar with the landscape of AI safety problems, including current approaches to alignment, robustness, and specification gaming.

You possess a strong track record of providing high quality developmental feedback on technical work, with examples of helping researchers improve their clarity and methodological precision.

You demonstrate excellent written and verbal communication skills, able to distill intricate technical concepts into structured narratives for diverse audiences.

You have hands on experience coordinating complex programs or projects, such as academic initiatives, research teams, or fellowship cohorts, with multiple moving parts.

You show consistent intellectual rigor, comfort with ambiguity, and a disciplined approach to updating beliefs based on new evidence and critical scrutiny.

You are mission driven regarding advanced AI risks, with a clear understanding of catastrophic scenarios and the role of empirical research in mitigation.

Nice to have

Prior experience working directly with AI safety researchers or participating in technical alignment communities.

Background in formal verification, mechanistic interpretability, or governance policy analysis relevant to high stakes AI systems.

Experience designing and running structured reading groups or workshops on technical AI safety topics.

Practical notes

This role is full time and based in Cambridge, Massachusetts.

Fellows are expected to commit to the full fellowship duration as defined by program cycles.

The position may involve travel to academic conferences and partner institutions on an as needed basis.

CBAI is committed to building a diverse team and welcomes applications from researchers who bring diverse perspectives to AI safety.