Senior Staff Software Engineer – Kernels

About the role

d-Matrix is building the next generation AI compute engine. As a Senior Staff Software Engineer – Kernels you will lead the development of software kernels that power AI hardware, working in a hybrid setting from our Belgrade office.

Key responsibilities

• Productize the software stack for the AI compute engine, delivering high‑performance kernels.
• Design, develop, enhance, and maintain kernels for next‑generation AI hardware architectures.
• Map computational graphs from AI frameworks onto hardware, optimizing for performance and resource usage.
• Collaborate with compiler experts to extend the compiler infrastructure and with hardware teams (mixed‑signal, DSP, CPU) to ensure tight software‑hardware integration.
• Scale software deliverables within tight development windows while balancing hardware‑software co‑design trade‑offs.

Required profile

• MS in Computer Engineering, Math, Physics or related field with 10+ years industry experience, or PhD with 1+ year experience.
• Strong understanding of computer architecture, data structures, system software and machine‑learning fundamentals.
• Proficiency in C/C++ and Python development on Linux, using standard development tools.
• Experience implementing algorithms for specialized hardware (FPGAs, DSPs, GPUs, AI accelerators) and using libraries such as CUDA.
• Hands‑on experience with common ML operators (GEMM, Convolution, BLAS, SIMD) and embedded SIMD processors like Tensilica.

Required skills

• C / C++
• Python
• Linux development environment
• CUDA
• FPGA, DSP, GPU, AI accelerator programming
• ML operators (GEMM, Convolution, BLAS, SIMD)
• Embedded SIMD processors (Tensilica)
• Familiarity with ML frameworks (TensorFlow, PyTorch) and compilers (MLIR, LLVM, TVM, Glow)