Staff Software Engineer, AI Safety & Abuse Detection

Anthropic; San Francisco (CA); Hybrid; USD 320,000 - 485,000; Full time

Benefits offered by this job

Equity donation matching

Generous vacation and parental leave

Flexible working hours

Office space in San Francisco

Job summary

Anthropic is seeking software engineers for the Safeguards team to build safety and oversight mechanisms for our AI systems. You will monitor models, detect unwanted behavior, and enforce our terms of service to ensure user well-being at scale.

Ideal candidates have a CS degree, strong Python and TypeScript skills, and the ability to work across the stack with excellent communication to explain complex concepts to non-technical stakeholders in a fast-growing, safety-minded environment.

Qualifications

• Bachelor's degree in Computer Science, Software Engineering or comparable experience
• Proficiency in Python and TypeScript
• Ability to work across the stack
• Strong communication skills and ability to explain complex technical concepts to non-technical stakeholders

Responsibilities

• Develop monitoring systems to detect unwanted behaviors from API partners and surface in dashboards for analysts
• Build abuse detection mechanisms and infrastructure
• Surface abuse patterns to research teams to harden models at training
• Build robust, multi-layered defenses for real-time safety improvements at scale

Skills

Python

TypeScript

Across the stack

Communication skills

Education

Bachelor's degree in Computer Science or related field

Job description

Anthropic is seeking software engineers for the Safeguards team to build safety and oversight mechanisms for our AI systems. You will monitor models, detect unwanted behavior, and enforce our terms of service to ensure user well-being at scale.

Ideal candidates have a CS degree, strong Python and TypeScript skills, and the ability to work across the stack with excellent communication to explain complex concepts to non-technical stakeholders in a fast-growing, safety-minded environment.