Software Engineer, ML Infrastructure

Position Overview

• Location Type: Hybrid
• Employment Type: Full-time
• Salary: $280,000 - $485,000 USD

Anthropic’s mission is to create reliable, interpretable, and steerable AI systems. We want AI to be safe and beneficial for our users and for society as a whole. Our team is a quickly growing group of committed researchers, engineers, policy experts, and business leaders working together to build beneficial AI systems. This role involves building large-scale ML systems from the ground up, focusing on creating safe, steerable, and trustworthy systems. As a Software Engineer, you'll contribute to various aspects of our code and infrastructure, including cluster reliability, throughput improvement, scientific experiments, and developer tooling. You're encouraged to write code with a strong understanding of the research context and its importance.

Requirements

• Significant software engineering experience
• Results-oriented with a bias towards flexibility and impact
• Ability to pick up slack, even if it goes outside the job description
• Enjoy pair programming (we love to pair!)
• Desire to learn more about machine learning research
• Concern for the societal impacts of your work

Responsibilities

• Optimizing the throughput of new attention mechanisms
• Comparing the compute efficiency of different Transformer variants
• Preparing datasets for model consumption
• Scaling distributed training jobs to thousands of GPUs
• Writing design documents for fault tolerance strategies
• Creating interactive visualizations of attention between tokens in a language model
• Contributing to the reliability and efficiency of our ML infrastructure

Technical Expertise (Strongly Preferred)

• Experience with high-performance, large-scale ML systems
• Familiarity with GPUs, Kubernetes, Pytorch, or OS internals
• Knowledge of language modeling with transformers and reinforcement learning
• Experience with large-scale ETL processes
• Expertise in security and privacy best practices
• Understanding of machine learning infrastructure (GPUs, TPUs, or Trainium) and supporting networking infrastructure (NCCL)
• Low-level systems knowledge (e.g., Linux kernel tuning and eBPF)

Application Instructions

• Deadline to apply: None. Applications will be reviewed on a rolling basis.
• Education: Bachelor’s degree in a related field or equivalent experience.
• Location-based hybrid policy: Currently, we expect [details incomplete - further information needed].

Company Information

Anthropic’s mission is to create reliable, interpretable, and steerable AI systems. We want AI to be safe and beneficial for our users and for society as a whole. Our team is a quickly growing group of committed researchers, engineers, policy experts, and business leaders working together to build beneficial AI systems. We encourage applicants to apply even if there isn't an immediately open role, as we continually evaluate for top talent. Reapplication is welcome, but limited to once per year.