Security Engineer / Research Engineer, RL Security

About the role

About Anthropic

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.

About the role

Reinforcement learning is how Claude learns to reason, write code, and act autonomously over long horizons. At frontier scale, an RL run is a large distributed system: training, sampling, and environment execution run concurrently across a large fleet of accelerators and hosts. As an engineer in RL Engineering, you'll work on keeping that system secure as it scales.

This is a hands-on engineering role. You'll work in the RL codebase and the infrastructure underneath it, alongside the engineers and researchers who build and use it. The system is large and changes as fast as the research does, so security work here has to keep up with that pace without costing performance or velocity.

We're hiring from two backgrounds: security engineers who want to work on large-scale ML systems, and ML systems or research engineers who want to work on security. Strong candidates have depth in one and real interest in the other.

Key responsibilities

-

Design, build, and operate security controls for RL training systems and the infrastructure they run on

-

Work across the stack (orchestration, environments, training, inference) wherever the work is needed

-

Review designs and code across the RL stack, find security problems, and fix them

-

Build shared libraries and tooling that other RL engineers and researchers build on

-

Build the observability needed to understand what these systems are doing

-

Work with researchers and performance engineers to make sure security changes preserve training correctness and performance

-

Write clear design documents for what you build

Minimum qualifications

-

Strong software engineering skills in Python and at least one systems language such as Rust, C++, or Go

-

Experience designing, building, and operating large-scale distributed systems or security systems in production

-

Strong understanding of security fundamentals

-

Working knowledge of Linux, containers, and cloud infrastructure

-

Experience debugging complex problems across many hosts and services

-

Ability to reason about tradeoffs between security, performance, and reliability

-

Strong written communication, including design documents

Preferred qualifications

-

Experience running ML training or inference infrastructure at scale