Software Engineer (AI Training Infrastructure)

Location: San Francisco, United States

Salary: $220K–$405K USD annually; equity offered. USD salary ranges apply only to U.S.-based positions.

Company

Perplexity builds AI-native products and state-of-the-art models.

What you will do

• Build distributed systems that orchestrate AI workloads across global-scale compute infrastructure.
• Develop large-scale reinforcement learning and agent evaluation platforms for training models and optimizing agent harnesses.
• Define user and agent interfaces for designing, managing, and launching experiments.
• Create tools that make experiment results legible to humans and agents, supporting experiment retrospectives and closed-loop autoresearch.
• Solve complex problems at the intersection of AI, distributed systems, and research acceleration.

Requirements

• 2+ years of experience developing and shipping distributed systems for compute-intensive workloads, with a focus on reliability, scalability, and performance.
• Proficiency in a general-purpose backend language such as Python, Rust, C++, or Go.
• Experience with cloud infrastructure and container orchestration, such as AWS and Kubernetes.
• Familiarity with reinforcement learning, agent evaluation, or machine learning infrastructure.
• Interest in building systems that accelerate experimentation and model development.
• BS, MS, or PhD in Computer Science, Engineering, or a related field, or equivalent practical experience.

Culture & Benefits

• Focus on craftsmanship and high-quality products for AI-native and AI-curious users.
• Ownership of identifying problems, designing solutions, and shipping them.
• Entrepreneurial environment with urgency and founder-like initiative.
• Opportunities to learn from talented peers and pursue knowledge and product improvement.
• Collaborative partnership across teams and disciplines.
• U.S. full-time employees receive equity, health, dental, vision, retirement, fitness, commuter, and dependent care benefits; international benefits are tailored to the region of residence.