Required Qualifications: Extensive experience in optimising AI chip architectures and AI systems, with deep familiarity with mainstream heterogeneous computing software and hardware architectures. Comprehensive expertise spanning applications, foundational software, and chip design. Hands-on experience in at least one of the following areas: numerical computation, compilation, algorithm and chip co-design, runtime systems, or shared memory management. Solid understanding of AI industry application scenarios, mainstream models, and algorithm development trends, with the ability to derive chip-layer requirements from these insights. Expertise in analysing workload sensitivity to micro-architecture features, evaluating performance trade-offs, and providing recommendations to optimise both micro-architecture and application software. Familiarity with the performance impact of various compute, memory, and communication configurations, as well as hardware and software implementation choices for AI acceleration. Proficiency with GPU compute APIs like CUDA or OpenCL, and experience leveraging GPU/NPU-optimised libraries to enhance performance. Practical experience in developing deep learning frameworks, compilers, or system software. Strong background in compiler optimisation techniques; familiarity with LLVM-MLIR is a plus. Proficiency in software development using C/C++ and Python. Desired Qualifications: Relevant experience in multiple subfields of AI, including application algorithms, frameworks, runtime systems, modelling and simulation, and compilers. In-depth understanding of innovative methods, platforms, and tools used by leading AI manufacturers, with proven experience in translating academic or research achievements into commercial products. Experience with GPU acceleration using AMD or NVIDIA GPUs. Expertise in developing inference backends and compilers for GPU or NPU systems. Proficiency with AI/ML inference frameworks such as ONNXRuntime, IREE, or TVM. Practical experience deploying AI models in production environments.
