Software Engineer, Accelerator Build Infrastructure

Details

Anthropic’s mission is to create reliable, interpretable, and steerable AI systems. The role is a systems-level engineering position focused on build infrastructure and low-level systems optimization, maintaining and improving non-trivial C/C++ builds and host-level systems. The position requires deep technical knowledge of compilation processes, hardware-software interfaces, build systems, and the ability to debug and optimize at the system level.

Responsibilities

• Maintain and improve complex build and packaging systems (Nix, pip, uv, CMake, Bazel, Make, etc.).
• Manage and interact with CI systems to support complex builds.
• Diagnose and resolve linking issues, symbol resolution problems, and toolchain/ABI incompatibilities.
• Debug C/C++ code and builds, including cross-compiling and embedded-system scenarios.
• Work with system calls, POSIX APIs, and kernel interfaces to debug and optimize runtime behavior.
• Use and extend toolchain debugging tools (readelf, bloaty, c++filt, nm, etc.) to investigate binary size, symbols, and object files.
• Configure and tune compiler flags and custom toolchains to achieve performance and compatibility goals.

Requirements

• 5+ years of experience in systems programming or infrastructure roles.
• Expert-level proficiency with build/packaging systems (experience with Nix is a major plus).
• Strong C/C++ debugging skills, including experience with cross compiling/linking and embedded systems is desirable.
• Familiarity with compilers and compilation pipelines (IR levels, compiler passes) and experience with cross-compilers.
• Understanding of linking processes, object file formats (ELF, DWARF), and ABI compatibility.
• Comfortable using toolchain and binary inspection tools such as readelf, bloaty, c++filt, and nm.
• Strong debugging mindset and ability to navigate large, legacy codebases.
• Often comes from backgrounds like HPC, game engine development, embedded systems, OS, or compiler teams.
• Bachelor's degree in a related field or equivalent experience (required) and eligibility for or successful sponsorship of visas may be supported on a case-by-case basis.

Preferred / Nice-to-have

• Nix experience (highlighted as a strong plus).
• Basic understanding of deep learning frameworks (PyTorch, Jax) from a systems perspective and familiarity with tensor operations.
• Experience with distributed training infrastructure.
• Experience diagnosing toolchain/ABI incompatibilities and resolving symbol resolution/linking issues.

Compensation & Logistics

• Expected base annual salary: $315,000 - $405,000 USD. Total compensation includes equity, benefits, and may include incentive compensation.
• Location-based hybrid policy: staff expected to be in one of Anthropic's offices at least ~25% of the time (some roles may require more time on-site).
• Visa sponsorship: Anthropic does sponsor visas in many cases and retains an immigration lawyer to assist where possible.

Why Anthropic / Culture

Anthropic values collaborative, high-impact AI research conducted as large-scale, cohesive efforts. The company emphasizes communication, frequent research discussions, and inclusion of diverse perspectives to build steerable, trustworthy AI systems.

How to Apply

Applications are reviewed on a rolling basis. Applicants are encouraged to apply even if they do not meet every qualification listed.