d-Matrix

Principal Software Engineer developing analytical performance models and architecture simulators for d-Matrix AI inference accelerators. The role bridges hardware and software teams to evaluate emerging workloads, project silicon performance, and guide HW/SW improvements.

Own and automate the infrastructure powering d-Matrix’s AI accelerator development, research, and customer environments across colocation, on-premises, and cloud platforms. This hands-on six-month contract role focuses on reliability, observability, infrastructure as code, and incident response, with potential conversion to full-time.

Build and optimize production-grade LLM inference systems across D-Matrix silicon, CPUs, GPUs, and custom accelerators. This principal-level role spans research prototypes, custom kernels, distributed serving, and hardware-aware performance optimization.

Join d-Matrix as a Senior Staff ML Researcher to innovate and optimize algorithms for large language model inference on advanced DNN accelerators.

Build and optimize software kernels and compiler infrastructure for d-Matrix’s next-generation AI compute hardware. This principal-level role combines deep computer architecture expertise with hands-on C/C++, Python, specialized-hardware, and ML software development.

Build and optimize runtime firmware, drivers, and systems software for d-Matrix’s multicore AI inference platform. This staff-level role owns runtime performance and collaborates closely with hardware and software teams in a hybrid Santa Clara environment.

Build and scale the system software and deployment infrastructure powering d-Matrix’s AI compute engine. This principal-level role combines distributed high-performance software, machine learning systems, and hardware-software co-design.

Lead the development and automation of software and machine-learning tests for d-Matrix’s next-generation AI compute engine. You’ll guide junior engineers while building Python-based test suites, regression infrastructure, and validation for AI chip software.