Alimama-Senior Expert in AI Serving

We are the Alimama engineering platform team of Taotian Group, committed to building a The industry's most powerful AI online service platform for searching promotion scenarios and AIGX applications. The team serves the core business of Taotian e-commerce and undertakes key responsibilities such as ultra-large-scale feature engineering of advertising models and the research and development and optimization of large model services such as AI inference and LLM/Diffusion. It is also responsible for high-performance online services (recommendation systems/multimodal large models) /Large language model), distributed systems (computing/storage/network), heterogeneous computing and AI compilation optimization (CPU/GPU/XPU) and other topics all present the most cutting-edge challenges in the industry. 1) Responsible for the design and implementation of the ultra-large-scale generative model online service system architecture 2) Responsible for the optimization of offline computing resources for large-scale heterogeneous hardware system inference 3) Responsible for the overall optimization of the model delivery pipeline. ,We hope that you: - Have a professional background in computer, software engineering, electronic information, automation, mathematics and other related fields. Doctoral degree is preferred. - Experience in the design and development of very large-scale distributed machine learning systems and large model online inference, understanding of cutting-edge technologies related to MLSys and the ability to promote innovative practices familiar with TensorFlow/PyTorch/XLA/MLIR/TVM/TRT-LLM/DeepSpeed/Megatron, etc. Industry mainstream AI frameworks or compilers are preferred. - Have first-line C++/Python engineering capabilities, be proficient in data structures and common algorithms, master various compilation, debugging, and performance analysis tools, have a CPU/GPU/ASIC system software or hardware background, be familiar with computer architecture, and be familiar with parallel programming (CUDA) /OpenCL, etc.) those with code generation and optimization background, familiar with the implementation of mainstream compilers such as GCC/LLVM will be given priority. - Computer, software engineering, electronic information, automation, mathematics and other related professional backgrounds, PhD is preferred. - Experience in the design and development of very large-scale distributed machine learning systems, familiar with large model online inference systems, understanding of cutting-edge technologies related to MLSys and the ability to promote innovative practices familiar with industry mainstream AI frameworks or compilers such as TensorFlow/PyTorch/XLA/MLIR/TVM priority. - Have first-line C++ engineering capabilities, be proficient in data structures and common algorithms, master various compilation, debugging, and performance analysis tools have a CPU/GPU/ASIC system software or hardware background, be familiar with computer architecture, and be familiar with parallel programming (CUDA/OpenCL) etc.), have a background in code generation and optimization, and be familiar with the implementation of mainstream compilers such as GCC/LLVM, etc. are preferred.