This list contains only the countries for which job offers have been published in the selected language (e.g., in the French version, only job offers written in French are displayed, and in the English version, only those in English).
We are seeking an experienced and highly technical SMTS Software Development Engineer to drive the design, development, optimization, and maintenance of the HIP runtime, a core component of the ROCm software stack. In this role, you will contribute to the evolution of HIP as a high‑performance C++ GPU programming environment, enabling developers to write portable, efficient GPU applications for AMD GPUs targeting data‑center, cloud, and AI workloads.
Job Responsibility:
Architect, implement, and optimize features in the HIP runtime, including memory management, kernel dispatch, device abstraction, multi‑GPU coordination, and synchronization primitives
Contribute to the evolution of the HIP programming model and interoperability with ROCr, HSA runtime, and compiler toolchains
Ensure functional correctness, performance, and scalability of runtime APIs across different GPU generations
Conduct root‑cause analysis and systems‑level debugging across the runtime, driver, compiler, and hardware layers
Profile GPU applications and internal runtime components to identify bottlenecks and design performance improvements
Optimize HIP runtime behavior for large-scale AI, HPC, and cloud workloads
Work closely with compiler teams (LLVM/Clang), driver teams, GPU architecture, and systems engineers to deliver end‑to‑end GPU software solutions
Contribute to API specifications and collaborate with upstream open-source communities where appropriate
Define and drive technical strategy for correctness, reliability, and conformance of the HIP runtime
Support enhancements in automated testing, CI, and stress/failure scenarios in the HIP test suite
Review design proposals, code reviews, and technical documents for quality and architectural alignment
Mentor junior engineers and contribute to establishing best practices in systems software development
Evaluate complex technical tradeoffs and drive decisions in runtime design and optimization
Represent the HIP runtime in technical discussions across ROCm teams and with strategic partners
Requirements:
Bachelor’s or Master’s degree in Computer Science, Computer Engineering, or related technical field
8+ years of software engineering experience in systems software, runtime libraries, GPU programming, or compiler/runtime interfaces
Strong proficiency in modern C++ (C++14/C++17 or newer), templates, memory models, and low‑level systems programming
Deep understanding of at least one GPU computing model (HIP, CUDA, SYCL, OpenCL, OpenMP offload)
Hands‑on experience with runtime systems, driver interfaces, or high‑performance compute libraries
Strong debugging skills using tools such as gdb, sanitizers, profilers, and GPU debugging tools