游戏AI-高性能推理系统研发专家
Tencent
游戏AI-高性能推理系统研发专家
Job Description
游戏AI-高性能推理系统研发专家
📋 Job Overview
The role focuses on developing high-performance inference systems for game AI within Tencent's 'Kaiwu' platform, supporting AIGC and LLM training/inference. It involves designing and optimizing GPU-based high-performance computing projects, including low-level performance tuning for chips like NVIDIA and AMD. The position emphasizes framework enhancements, hardware analysis, and exploring cutting-edge AI technologies to boost efficiency in AI engineering.
📍 Location: Shenzhen, China
🏢 Business Unit: TEG
📄 Full Description
1.负责腾讯“开悟”平台的AIGC及LLM的训练/推理支持，参与基于GPU的高性能计算（HPC）项目设计与开发，负责GPU芯片（NVIDIA/AMD等）的底层性能优化与调优；
2.针对大模型推理、训练等场景，优化和扩展vLLM、SGLang等框架的核心模块，提升计算效率与资源利用率；
3.深入分析GPU硬件架构特性（如Tensor Core、显存带宽、通信机制等），设计并实现高性能算子与算法；
4.探索前沿技术方向（如混合专家模型MoE、动态计算图优化等），推动AI工程化落地的效率提升。
2.针对大模型推理、训练等场景，优化和扩展vLLM、SGLang等框架的核心模块，提升计算效率与资源利用率；
3.深入分析GPU硬件架构特性（如Tensor Core、显存带宽、通信机制等），设计并实现高性能算子与算法；
4.探索前沿技术方向（如混合专家模型MoE、动态计算图优化等），推动AI工程化落地的效率提升。
🎯 Key Responsibilities
- Responsible for training/inference support of AIGC and LLM on Tencent's 'Kaiwu' platform, participating in GPU-based high-performance computing (HPC) project design and development, and handling low-level performance optimization and tuning for GPU chips (NVIDIA/AMD etc.).
- For large model inference and training scenarios, optimize and extend core modules of frameworks like vLLM and SGLang to improve computational efficiency and resource utilization.
- Deeply analyze GPU hardware architecture features (such as Tensor Core, memory bandwidth, communication mechanisms, etc.), and design and implement high-performance operators and algorithms.
- Explore cutting-edge technical directions (such as Mixture of Experts MoE, dynamic computation graph optimization, etc.) to drive efficiency improvements in AI engineering implementation.
🛠️ Required Skills
- GPU programming and optimization (NVIDIA/AMD chips)
- High-performance computing (HPC) design and development
- Framework optimization (vLLM, SGLang)
- GPU hardware architecture analysis (Tensor Core, memory bandwidth, communication)
- High-performance operators and algorithms implementation
- Knowledge of advanced AI technologies (MoE, dynamic computation graph optimization)
Locations
- Shenzhen, China
Salary
Estimated Salary Rangemedium confidence
800,000 - 1,500,000 CNY / yearly
Source: ai estimated
* This is an estimated range based on market data and may vary based on experience and qualifications.
Skills Required
- GPU programming and optimization (NVIDIA/AMD chips)intermediate
- High-performance computing (HPC) design and developmentintermediate
- Framework optimization (vLLM, SGLang)intermediate
- GPU hardware architecture analysis (Tensor Core, memory bandwidth, communication)intermediate
- High-performance operators and algorithms implementationintermediate
- Knowledge of advanced AI technologies (MoE, dynamic computation graph optimization)intermediate
Responsibilities
- Responsible for training/inference support of AIGC and LLM on Tencent's 'Kaiwu' platform, participating in GPU-based high-performance computing (HPC) project design and development, and handling low-level performance optimization and tuning for GPU chips (NVIDIA/AMD etc.).
- For large model inference and training scenarios, optimize and extend core modules of frameworks like vLLM and SGLang to improve computational efficiency and resource utilization.
- Deeply analyze GPU hardware architecture features (such as Tensor Core, memory bandwidth, communication mechanisms, etc.), and design and implement high-performance operators and algorithms.
- Explore cutting-edge technical directions (such as Mixture of Experts MoE, dynamic computation graph optimization, etc.) to drive efficiency improvements in AI engineering implementation.
Tags & Categories
Answer 10 quick questions to check your fit for 游戏AI-高性能推理系统研发专家 @ Tencent.
10 Questions
~2 Minutes
Instant Score
Related Books and Jobs
游戏AI-高性能推理系统研发专家
Tencent
游戏AI-高性能推理系统研发专家
Job Description
游戏AI-高性能推理系统研发专家
📋 Job Overview
The role focuses on developing high-performance inference systems for game AI within Tencent's 'Kaiwu' platform, supporting AIGC and LLM training/inference. It involves designing and optimizing GPU-based high-performance computing projects, including low-level performance tuning for chips like NVIDIA and AMD. The position emphasizes framework enhancements, hardware analysis, and exploring cutting-edge AI technologies to boost efficiency in AI engineering.
📍 Location: Shenzhen, China
🏢 Business Unit: TEG
📄 Full Description
1.负责腾讯“开悟”平台的AIGC及LLM的训练/推理支持，参与基于GPU的高性能计算（HPC）项目设计与开发，负责GPU芯片（NVIDIA/AMD等）的底层性能优化与调优；
2.针对大模型推理、训练等场景，优化和扩展vLLM、SGLang等框架的核心模块，提升计算效率与资源利用率；
3.深入分析GPU硬件架构特性（如Tensor Core、显存带宽、通信机制等），设计并实现高性能算子与算法；
4.探索前沿技术方向（如混合专家模型MoE、动态计算图优化等），推动AI工程化落地的效率提升。
2.针对大模型推理、训练等场景，优化和扩展vLLM、SGLang等框架的核心模块，提升计算效率与资源利用率；
3.深入分析GPU硬件架构特性（如Tensor Core、显存带宽、通信机制等），设计并实现高性能算子与算法；
4.探索前沿技术方向（如混合专家模型MoE、动态计算图优化等），推动AI工程化落地的效率提升。
🎯 Key Responsibilities
- Responsible for training/inference support of AIGC and LLM on Tencent's 'Kaiwu' platform, participating in GPU-based high-performance computing (HPC) project design and development, and handling low-level performance optimization and tuning for GPU chips (NVIDIA/AMD etc.).
- For large model inference and training scenarios, optimize and extend core modules of frameworks like vLLM and SGLang to improve computational efficiency and resource utilization.
- Deeply analyze GPU hardware architecture features (such as Tensor Core, memory bandwidth, communication mechanisms, etc.), and design and implement high-performance operators and algorithms.
- Explore cutting-edge technical directions (such as Mixture of Experts MoE, dynamic computation graph optimization, etc.) to drive efficiency improvements in AI engineering implementation.
🛠️ Required Skills
- GPU programming and optimization (NVIDIA/AMD chips)
- High-performance computing (HPC) design and development
- Framework optimization (vLLM, SGLang)
- GPU hardware architecture analysis (Tensor Core, memory bandwidth, communication)
- High-performance operators and algorithms implementation
- Knowledge of advanced AI technologies (MoE, dynamic computation graph optimization)
Locations
- Shenzhen, China
Salary
Estimated Salary Rangemedium confidence
800,000 - 1,500,000 CNY / yearly
Source: ai estimated
* This is an estimated range based on market data and may vary based on experience and qualifications.
Skills Required
- GPU programming and optimization (NVIDIA/AMD chips)intermediate
- High-performance computing (HPC) design and developmentintermediate
- Framework optimization (vLLM, SGLang)intermediate
- GPU hardware architecture analysis (Tensor Core, memory bandwidth, communication)intermediate
- High-performance operators and algorithms implementationintermediate
- Knowledge of advanced AI technologies (MoE, dynamic computation graph optimization)intermediate
Responsibilities
- Responsible for training/inference support of AIGC and LLM on Tencent's 'Kaiwu' platform, participating in GPU-based high-performance computing (HPC) project design and development, and handling low-level performance optimization and tuning for GPU chips (NVIDIA/AMD etc.).
- For large model inference and training scenarios, optimize and extend core modules of frameworks like vLLM and SGLang to improve computational efficiency and resource utilization.
- Deeply analyze GPU hardware architecture features (such as Tensor Core, memory bandwidth, communication mechanisms, etc.), and design and implement high-performance operators and algorithms.
- Explore cutting-edge technical directions (such as Mixture of Experts MoE, dynamic computation graph optimization, etc.) to drive efficiency improvements in AI engineering implementation.
Tags & Categories
Answer 10 quick questions to check your fit for 游戏AI-高性能推理系统研发专家 @ Tencent.
10 Questions
~2 Minutes
Instant Score
Related Books and Jobs