Product Safety Engineer, Multimodal AI

SpaceXSI; Palo Alto (CA); On-site; USD 180,000 - 440,000; Full time

Benefits offered by this job

Equity

Comprehensive medical, vision, and d 6

401(k) retirement plan

Short & long-term disability insurance

Life insurance

Various discounts and perks

Job summary

SpaceXSI is seeking a Member of Technical Staff - Imagine Safety to design scalable safety systems for Grok's multimodal generation platform, ensuring responsible deployment at scale and protecting users.

You will work across engineering, research, and policy teams to iterate safety features, measure risk, and maintain high performance with low latency while advancing humanity’s understanding.

Qualifications

• Proficiency in Rust with a track record of scalable production code.
• Experience building safety or content moderation systems at scale.
• Hands-on with real-time detection, data pipelines, or evaluation frameworks.
• Ability to deliver robust, reliable solutions for millions of users.
• Strong problem-solving and responsible AI focus.

Responsibilities

• Design and implement scalable safety systems for Grok's media platform, including real-time moderation and safeguards.
• Build infrastructure to measure and mitigate safety risks at global scale.
• Develop tools and evaluation frameworks in collaboration with researchers, product, and policy teams.
• Create feedback loops between user interactions, model outputs, and training data.
• Own full-cycle safety feature development from prototyping to deployment and monitoring.
• Partner with Imagine teams to embed safety into core features.

Skills

Rust

Safety systems design

Real-time detection

Production systems

High uptime & performance

Problem solving

Responsible AI

Job description

SpaceXSI is seeking a Member of Technical Staff - Imagine Safety to design scalable safety systems for Grok's multimodal generation platform, ensuring responsible deployment at scale and protecting users.

You will work across engineering, research, and policy teams to iterate safety features, measure risk, and maintain high performance with low latency while advancing humanity’s understanding.