Resume and JobRESUME AND JOB
Tencent logo

高性能计算工程师-国产化方向

Tencent

高性能计算工程师-国产化方向

Tencent logo

Tencent

full-time

Posted: December 11, 2025

Number of Vacancies: 1

Job Description

高性能计算工程师-国产化方向

📋 Job Overview

Tencent is seeking a High-Performance Computing Engineer specializing in domestic chip technologies to optimize AI models and frameworks for Chinese hardware platforms. The role involves adapting algorithms to chips like Ascend, Hygon, Cambricon, and Kunlun, while enhancing performance in large model inference scenarios. Responsibilities include developing high-performance operators, optimizing distributed systems, and ensuring efficient execution on domestic architectures.

📍 Location: Xi'an, China

🏢 Business Unit: CSIG

📄 Full Description

1.产硬件适配与优化开发​:参与基于昇腾、海光、寒武纪思元、昆仑芯等国产化芯片的算法模型适配,负责底层性能调优,针对不同芯片架构特性制定差异化优化方案;
2.国产框架与引擎优化​:针对文生文、生图、生视频等大模型推理场景,扩展 vLLM/SGLang 等主流框架的国产化硬件支持能力,重点提升 FP8 精度模型适配、KV Cache 国产化存储优化等关键场景效率;
3.国产化算子与算法研发​:深入剖析国产芯片架构特性,设计实现高性能算子库,重点突破 Matmul、MoE 等核心算子的指令级优化,确保精度与性能平衡;
4.分布式系统协同优化​:解决国产芯片集群下模型并行、数据并行的性能瓶颈,优化多卡互联通信机制,提升大模型的分布式运行效率。​。

🎯 Key Responsibilities

  • 参与基于昇腾、海光、寒武纪思元、昆仑芯等国产化芯片的算法模型适配,负责底层性能调优,针对不同芯片架构特性制定差异化优化方案
  • 针对文生文、生图、生视频等大模型推理场景,扩展 vLLM/SGLang 等主流框架的国产化硬件支持能力,重点提升 FP8 精度模型适配、KV Cache 国产化存储优化等关键场景效率
  • 深入剖析国产芯片架构特性,设计实现高性能算子库,重点突破 Matmul、MoE 等核心算子的指令级优化,确保精度与性能平衡
  • 解决国产芯片集群下模型并行、数据并行的性能瓶颈,优化多卡互联通信机制,提升大模型的分布式运行效率

🛠️ Required Skills

  • 国产化芯片适配与优化 (e.g., Ascend, Hygon, Cambricon, Kunlun)
  • 大模型推理框架扩展 (e.g., vLLM, SGLang)
  • 高性能算子研发 (e.g., Matmul, MoE)
  • 分布式系统优化 (模型并行、数据并行、多卡通信)

Locations

  • Xi'an, China

Salary

Estimated Salary Rangemedium confidence

300,000 - 600,000 CNY / yearly

Source: ai estimated

* This is an estimated range based on market data and may vary based on experience and qualifications.

Skills Required

  • 国产化芯片适配与优化 (e.g., Ascend, Hygon, Cambricon, Kunlun)intermediate
  • 大模型推理框架扩展 (e.g., vLLM, SGLang)intermediate
  • 高性能算子研发 (e.g., Matmul, MoE)intermediate
  • 分布式系统优化 (模型并行、数据并行、多卡通信)intermediate

Responsibilities

  • 参与基于昇腾、海光、寒武纪思元、昆仑芯等国产化芯片的算法模型适配,负责底层性能调优,针对不同芯片架构特性制定差异化优化方案
  • 针对文生文、生图、生视频等大模型推理场景,扩展 vLLM/SGLang 等主流框架的国产化硬件支持能力,重点提升 FP8 精度模型适配、KV Cache 国产化存储优化等关键场景效率
  • 深入剖析国产芯片架构特性,设计实现高性能算子库,重点突破 Matmul、MoE 等核心算子的指令级优化,确保精度与性能平衡
  • 解决国产芯片集群下模型并行、数据并行的性能瓶颈,优化多卡互联通信机制,提升大模型的分布式运行效率

Target Your Resume for "高性能计算工程师-国产化方向" , Tencent

Get personalized recommendations to optimize your resume specifically for 高性能计算工程师-国产化方向. Takes only 15 seconds!

AI-powered keyword optimization
Skills matching & gap analysis
Experience alignment suggestions

Check Your ATS Score for "高性能计算工程师-国产化方向" , Tencent

Find out how well your resume matches this job's requirements. Get comprehensive analysis including ATS compatibility, keyword matching, skill gaps, and personalized recommendations.

ATS compatibility check
Keyword optimization analysis
Skill matching & gap identification
Format & readability score

Tags & Categories

TencentXi'anChinaCSIGCSIG

Related Jobs You May Like

No related jobs found at the moment.

Tencent logo

高性能计算工程师-国产化方向

Tencent

高性能计算工程师-国产化方向

Tencent logo

Tencent

full-time

Posted: December 11, 2025

Number of Vacancies: 1

Job Description

高性能计算工程师-国产化方向

📋 Job Overview

Tencent is seeking a High-Performance Computing Engineer specializing in domestic chip technologies to optimize AI models and frameworks for Chinese hardware platforms. The role involves adapting algorithms to chips like Ascend, Hygon, Cambricon, and Kunlun, while enhancing performance in large model inference scenarios. Responsibilities include developing high-performance operators, optimizing distributed systems, and ensuring efficient execution on domestic architectures.

📍 Location: Xi'an, China

🏢 Business Unit: CSIG

📄 Full Description

1.产硬件适配与优化开发​:参与基于昇腾、海光、寒武纪思元、昆仑芯等国产化芯片的算法模型适配,负责底层性能调优,针对不同芯片架构特性制定差异化优化方案;
2.国产框架与引擎优化​:针对文生文、生图、生视频等大模型推理场景,扩展 vLLM/SGLang 等主流框架的国产化硬件支持能力,重点提升 FP8 精度模型适配、KV Cache 国产化存储优化等关键场景效率;
3.国产化算子与算法研发​:深入剖析国产芯片架构特性,设计实现高性能算子库,重点突破 Matmul、MoE 等核心算子的指令级优化,确保精度与性能平衡;
4.分布式系统协同优化​:解决国产芯片集群下模型并行、数据并行的性能瓶颈,优化多卡互联通信机制,提升大模型的分布式运行效率。​。

🎯 Key Responsibilities

  • 参与基于昇腾、海光、寒武纪思元、昆仑芯等国产化芯片的算法模型适配,负责底层性能调优,针对不同芯片架构特性制定差异化优化方案
  • 针对文生文、生图、生视频等大模型推理场景,扩展 vLLM/SGLang 等主流框架的国产化硬件支持能力,重点提升 FP8 精度模型适配、KV Cache 国产化存储优化等关键场景效率
  • 深入剖析国产芯片架构特性,设计实现高性能算子库,重点突破 Matmul、MoE 等核心算子的指令级优化,确保精度与性能平衡
  • 解决国产芯片集群下模型并行、数据并行的性能瓶颈,优化多卡互联通信机制,提升大模型的分布式运行效率

🛠️ Required Skills

  • 国产化芯片适配与优化 (e.g., Ascend, Hygon, Cambricon, Kunlun)
  • 大模型推理框架扩展 (e.g., vLLM, SGLang)
  • 高性能算子研发 (e.g., Matmul, MoE)
  • 分布式系统优化 (模型并行、数据并行、多卡通信)

Locations

  • Xi'an, China

Salary

Estimated Salary Rangemedium confidence

300,000 - 600,000 CNY / yearly

Source: ai estimated

* This is an estimated range based on market data and may vary based on experience and qualifications.

Skills Required

  • 国产化芯片适配与优化 (e.g., Ascend, Hygon, Cambricon, Kunlun)intermediate
  • 大模型推理框架扩展 (e.g., vLLM, SGLang)intermediate
  • 高性能算子研发 (e.g., Matmul, MoE)intermediate
  • 分布式系统优化 (模型并行、数据并行、多卡通信)intermediate

Responsibilities

  • 参与基于昇腾、海光、寒武纪思元、昆仑芯等国产化芯片的算法模型适配,负责底层性能调优,针对不同芯片架构特性制定差异化优化方案
  • 针对文生文、生图、生视频等大模型推理场景,扩展 vLLM/SGLang 等主流框架的国产化硬件支持能力,重点提升 FP8 精度模型适配、KV Cache 国产化存储优化等关键场景效率
  • 深入剖析国产芯片架构特性,设计实现高性能算子库,重点突破 Matmul、MoE 等核心算子的指令级优化,确保精度与性能平衡
  • 解决国产芯片集群下模型并行、数据并行的性能瓶颈,优化多卡互联通信机制,提升大模型的分布式运行效率

Target Your Resume for "高性能计算工程师-国产化方向" , Tencent

Get personalized recommendations to optimize your resume specifically for 高性能计算工程师-国产化方向. Takes only 15 seconds!

AI-powered keyword optimization
Skills matching & gap analysis
Experience alignment suggestions

Check Your ATS Score for "高性能计算工程师-国产化方向" , Tencent

Find out how well your resume matches this job's requirements. Get comprehensive analysis including ATS compatibility, keyword matching, skill gaps, and personalized recommendations.

ATS compatibility check
Keyword optimization analysis
Skill matching & gap identification
Format & readability score

Tags & Categories

TencentXi'anChinaCSIGCSIG

Related Jobs You May Like

No related jobs found at the moment.