Interpretability Jobs

Anthropic

Anthropic is seeking individuals for a paid, fixed-term research residency program, lasting between six weeks and six months. This program is designed for experienced professionals, such as professors in quantitative disciplines, technical founders, or staff-level engineers, who are looking to transition into full-time work focused on ensuring AI systems are safe and beneficial for humanity. While a background in machine learning is not strictly required, candidates should be motivated by AI safety and excited about empirical AI research. The residency offers a scoped project with direct mentorship from Anthropic researchers, access to Claude models and GPU compute, and connections to the AI safety and security research community.

$350k - $500k

Raindrop is seeking a founding Research Engineer to investigate the behavior of AI agents in production environments. This role is crucial for an early-stage team, involving setting the research direction, conducting experiments, and disseminating findings through publications. The focus will be on understanding agent failures, advancing simulation techniques, developing novel detection methods for undefined issues, and exploring high-impact open questions in AI agent behavior. The ideal candidate will possess strong communication skills, a passion for sharing research, deep expertise in areas like interpretability or AI safety, and the ability to prototype and ship experiments.

Scale AI

Scale Labs is seeking a Research Scientist focused on Safety Post-Training to develop and apply methods for making frontier AI systems safer and more understandable. This role involves designing and executing post-training pipelines to analyze how training choices impact model safety, robustness, and alignment. You will also develop evaluations informed by interpretability techniques to uncover the reasons behind undesirable model behaviors and use these insights to guide targeted safety mitigations. A key aspect of this position is collaborating with policymakers, engineers, and other researchers to translate findings into actionable safety standards, benchmarks, and best practices.

$216k - $270k

Anthropic

Anthropic's mission is to create reliable, interpretable, and steerable AI systems that are safe and beneficial for users and society. The Interpretability team focuses on mechanistic interpretability, aiming to discover how neural network parameters map to meaningful algorithms. This involves reverse-engineering how trained models work, treating neural networks like biological systems or computer programs to be reverse-engineered. The team's goal is to create a solid scientific foundation for understanding neural networks and making them safe, with a focus on issues like "superposition" and decomposing models into interpretable components. This role supports a team of expert researchers and engineers working on these ambitious safety research goals.

$350k - $500k

Anthropic

The Interpretability team at Anthropic is dedicated to reverse-engineering how trained models work, believing that a mechanistic understanding is crucial for making advanced AI systems safe and steerable. This role focuses on mechanistic interpretability, aiming to discover how neural network parameters map to meaningful algorithms, akin to performing "biology" or "neuroscience" on neural networks. The team has made significant progress in understanding issues like "superposition" and decomposing models into interpretable components, with recent work identifying millions of features in production models and developing methods to understand multi-hop reasoning and planning. This research is foundational for building reliable, interpretable, and steerable AI systems and often involves collaboration with other teams at Anthropic to enhance AI safety.

$350k - $850k

CHAOS Industries is seeking a mission-oriented Applied AI Engineer to develop, integrate, and deploy AI/ML capabilities across its defense product lines. This role involves working with data scientists, software engineers, and product teams to conduct advanced AI research and translate it into reliable, real-world software, particularly for defense applications requiring extreme accuracy under challenging conditions. The ideal candidate possesses strong data science and software engineering fundamentals, a track record of technical excellence, and experience applying AI/ML to real products, with a preference for backgrounds in aerospace, defense, or similar mission-driven environments. Comfort with independent operation, rapid learning of new technical domains, cross-team collaboration, and moving prototypes to production with minimal oversight is essential.

$150k - $170k