14 days ago
Base Salary
$230k - $285k/yr
Responsibilities
- Design, build, and maintain functional, instruction-level, and performance-level software simulators and models for an AI accelerator.
- Build golden reference models, co-simulation, trace generation, and validation harnesses for compiler and runtime development.
- Improve simulator throughput using multithreading, JIT or binary translation, and distributed execution.
- Correlate models against RTL, emulation, and first silicon, and incorporate discrepancies into the models.
- Develop AI-assisted tool flows to accelerate model development and validation.
Requirements
- Exceptional modern C++ for performance-critical systems software and strong Python for tooling and infrastructure.
- At least 5 years of experience building simulators, emulators, virtual platforms, instruction-set simulators, or large-scale systems and infrastructure software.
- Strong computer-architecture fundamentals, including pipelines, memory hierarchy, ISAs, and accelerator or GPU-class datapaths.
- Experience making functional versus cycle-approximate modeling trade-offs, validating against references, and optimizing simulator performance.
- Strong software-engineering practices involving testing, CI, profiling, and debugging large, long-lived codebases.
- Optional experience with gem5, QEMU, SystemC-TLM, Spike, custom simulators, JIT or dynamic binary translation, ML frameworks and workloads, compiler/runtime systems, distributed simulation, or hardware/software co-design.
Benefits
- Equity grant according to company guidelines.
- Medical, dental, and vision benefits.
- 401(k) and standard paid time off.
- DensityAI sponsors qualified candidates for H-1B, O-1, TN, E-3, and other employment-based visas, with immigration support to secure or transfer status.
- Work authorization is required at the start of employment.
- The role may involve U.S. export-controlled information under EAR/ITAR, with licensing or scope adjustments discussed during the interview.
