Find your next AI-scored role.

Research Scientist/Engineer at Anthropic focused on honesty and alignment. Develop techniques to minimize hallucinations, enhance truthfulness in language models, and create evaluation frameworks for model accuracy and calibration.

Software engineer at Anthropic building safety and oversight mechanisms for AI systems. Focus on monitoring models, detecting abuse, and preventing misuse through detection systems and real-time defenses.

Software engineer at Anthropic building account abuse detection and enforcement systems. Focus on signal gathering, monitoring, and multi-layered defenses against bad actors using computing resources.

Research Lead on Anthropic's Training Insights team, designing and leading novel evaluation methodologies to measure model capabilities across training and deployment. Hands-on leadership role spanning model development lifecycle, with cross-organizational impact on safety and capability measurement.

Research Engineer in Reinforcement Learning at Anthropic, building scalable RL infrastructure and training methodologies for large language models. Focus on agentic systems, tool use, code generation, and reasoning capabilities.

Policy Manager at Anthropic designing evaluation methodologies and safety strategies for AI systems handling chemical weapons and explosives information. Requires Ph.D. in Chemistry or Chemical Engineering with 5-8+ years in C/E defense and expertise in energetic materials.

Offensive Security Research Engineer at Anthropic focused on safeguarding LLMs from adversarial misuse. Role involves vulnerability research, exploitation scaffolds, and defensive strategy development to mitigate risks from AI-enabled attacks.

Developer Education Lead at Anthropic, owning end-to-end strategy and execution for developer-facing education content across video, tutorials, and courses. Role sits at intersection of DevRel, product, and education, working cross-functionally on launches and content roadmap prioritization.

Software engineer at Anthropic building data collection pipelines and interfaces for model training. Focus on full-stack infrastructure, crowdworker experience, and supporting research teams' evolving data needs.

Safeguards Analyst at Anthropic building account abuse detection and enforcement systems. Focus on developing signals, integrating third-party data, and operationalizing enforcement workflows to protect the platform at scale.

Engineering Manager at Anthropic leading the Safeguards Data Infrastructure team. Responsible for building privacy-preserving data infrastructure, managing multi-cloud portability, and ensuring compliance with regulations like HIPAA across deployment environments.

Threat Collections Engineer at Anthropic building infrastructure for threat discovery and abuse detection. Develop automated detection systems, YARA rule infrastructure, and data pipelines integrating external threat intelligence sources.

Software engineer at Anthropic building safeguards infrastructure for AI systems. Focus on detection systems, real-time safety mechanisms, and operational tooling to prevent misuse and ensure user well-being.

Research Scientist at Anthropic's Frontier Red Team focusing on emerging societal risks from advanced AI systems. Design experiments, build evaluations, and develop defenses against risks from autonomous AI integration into business and infrastructure.

Design Engineer at Anthropic's Education Labs building product features that help users develop real skill with AI. Full-stack role combining research insights, interaction design, and production engineering to ship capability-focused experiences.

Technical CBRN-E Threat Investigator at Anthropic detecting and investigating misuse of AI systems for chemical, biological, radiological, nuclear, and explosives threats. Role combines AI safety with biosecurity expertise, conducting threat analysis and developing detection capabilities.

Research Scientist at Anthropic analyzing Claude's real-world usage patterns and building evaluations to assess model safety and alignment. Role involves cross-functional collaboration with fine-tuning, safeguards, and policy teams to translate research insights into model improvements.

Research Engineer at Anthropic building vision and spatial reasoning capabilities for Claude. Focus on multimodal model development, evaluation, and agentic infrastructure across pretraining, RL, and runtime techniques.

Research Engineer at Anthropic building autonomous AI systems and defensive agents to understand and counter adversarial AI. Focus on agent design, evals, robotics integration, and policy-relevant demonstrations.

Technical Policy Manager, Cyber Harms at Anthropic leading a team to design safety systems that detect harmful cyber behaviors and prevent AI misuse. Combines deep cybersecurity expertise with policy development to shape responsible AI safety in the cybersecurity domain.

Technical Cyber Threat Investigator at Anthropic detecting and investigating misuse of Claude for malicious cyber operations. Role combines AI safety and cybersecurity, developing detection techniques and building defenses against AI-enabled threats.

Research Engineer at Anthropic building training environments for agentic AI systems. Role blends research and engineering, focusing on reinforcement learning, environment design, and capability evaluation for safe AI.

Research Engineer at Anthropic building reinforcement learning systems for secure coding and cybersecurity. Role blends research and engineering, requiring domain expertise in cybersecurity and machine learning to develop RL environments and advance model capabilities in defensive security.

Biological Safety Research Scientist at Anthropic designing safety systems to detect harmful AI behaviors and prevent misuse of biological knowledge. Role involves capability evaluations, threat modeling, and embedding biosecurity safeguards throughout model development.

Senior Research Scientist at Anthropic leading research on reward models and RLHF for large language models. Focus on novel architectures, LLM-based evaluation methods, and techniques to mitigate reward hacking.