Hardware Tools Engineer

OpenAI OpenAI · AI Frontier · San Francisco, CA · Scaling

Develop and evolve the tooling ecosystem for hardware engineers at an AI company, focusing on compilers, IR transformations, simulation, debugging, and automation infrastructure for AI-native silicon.

What you'd actually do

  1. Build and improve the software tooling that makes hardware teams faster: compilation, IR transforms, RTL generation, simulation, debug, and automation.
  2. Extend and integrate hardware compiler stacks (frontends, IR passes, lowering, scheduling, codegen to Verilog/SystemVerilog) and connect them to real design workflows.
  3. Improve developer experience and reliability: reproducible builds, better error messages, faster iteration loops, and dependable CI and regression infrastructure.
  4. Work closely with designers and verification engineers to turn real pain points into durable tools.
  5. Dive into RTL when needed: read and reason about Verilog/SystemVerilog to debug issues, validate tool output, and improve debuggability.

Skills

Required

  • Demonstrated ability to build and maintain software (projects, internships, research, open source, or equivalent experience).
  • Strong CS fundamentals: data structures, algorithms, debugging, and software design.
  • Proficiency in at least one of Rust, C++, or Python (and willingness to learn the rest).
  • Familiarity with digital design concepts and the ability to read RTL (Verilog/SystemVerilog) or equivalent hardware descriptions.
  • Familiarity with compiler or IR-based ideas (representations, passes, transformations, lowering), through coursework or projects.
  • Comfort operating in ambiguity and iterating quickly with users of your tools.

Nice to have

  • Exposure to compiler and hardware toolchains such as XLS/DSLX, LLVM, Chisel/FIRRTL, CIRCT/MLIR, other novel hardware languages (e.g. HardCaml, SpinalHDL, Spade, PyMTL, Clash, BlueSpec, PyRope)
  • Experience with Verilog tooling ecosystems (Yosys/RTLIL, Verilator, Slang) or writing tooling around them.
  • Experience with build and test infrastructure (Bazel, CI systems, fuzzing, performance testing).
  • Prior work touching synthesis, place and route, static timing analysis, or other PPA-related workflows.

What the JD emphasized

  • direct impact on how quickly and effectively we design next-generation AI systems