元宝-大模型训练工程师
Tencent
元宝-大模型训练工程师
Job Description
元宝-大模型训练工程师
📋 Job Overview
The Yuanbao Large Model Training Engineer role at Tencent involves designing and optimizing LLM training frameworks to support high-performance training of large-scale models. The position requires close collaboration with algorithm and platform teams to ensure stable, high-performance, and scalable LLM infrastructure. Key focuses include system design, performance tuning, stability assurance, and driving AI large model innovations for business applications.
📍 Location: Beijing, China
🏢 Business Unit: CSIG
📄 Full Description
1.负责设计和优化LLM训练框架，支撑LLM高性能训练。与算法团队、平台团队紧密协作，确保LLM infra的稳定、高性能、可扩展，推动AI大模型技术的落地与创新：；
2.负责设计和优化LLM训练框架，支撑LLM高性能训练。与算法团队、平台团队紧密协作，确保LLM infra的稳定、高性能、可扩展；
3.系统设计与优化：设计并搭建分布式训练框架，跟平台协作，支持千亿级参数大模型的训练；
4.性能调优与成本优化：针对大模型训练任务，优化框架（如PyTorch/VERL等）的分布式策略，提升训练效率；
5.稳定性与可靠性保障：设计高可用架构，解决训练中断、数据丢失等风险，确保长周期训练任务的稳定性；
6.协作与落地：与算法团队紧密合作，理解模型需求，提供基础设施层面的技术建议；推动开源工具的定制化开发，适配业务场景。
2.负责设计和优化LLM训练框架，支撑LLM高性能训练。与算法团队、平台团队紧密协作，确保LLM infra的稳定、高性能、可扩展；
3.系统设计与优化：设计并搭建分布式训练框架，跟平台协作，支持千亿级参数大模型的训练；
4.性能调优与成本优化：针对大模型训练任务，优化框架（如PyTorch/VERL等）的分布式策略，提升训练效率；
5.稳定性与可靠性保障：设计高可用架构，解决训练中断、数据丢失等风险，确保长周期训练任务的稳定性；
6.协作与落地：与算法团队紧密合作，理解模型需求，提供基础设施层面的技术建议；推动开源工具的定制化开发，适配业务场景。
🎯 Key Responsibilities
- Design and optimize LLM training frameworks to support high-performance training, collaborating closely with algorithm and platform teams to ensure stable, high-performance, and scalable LLM infrastructure.
- Design and build distributed training frameworks in collaboration with the platform team, supporting training of billion-parameter large models.
- Perform performance tuning and cost optimization for large model training tasks, optimizing distributed strategies in frameworks like PyTorch/VERL to improve training efficiency.
- Ensure stability and reliability by designing high-availability architectures, addressing risks such as training interruptions and data loss for long-cycle training tasks.
- Collaborate with the algorithm team to understand model requirements, provide technical advice on infrastructure, and drive customized development of open-source tools to fit business scenarios.
🛠️ Required Skills
- Expertise in designing and optimizing distributed training frameworks for LLMs.
- Proficiency in performance tuning and optimization of frameworks like PyTorch or VERL.
- Knowledge of high-availability architectures and stability measures for large-scale training.
- Strong collaboration skills with algorithm and platform teams.
- Experience in pushing AI large model technologies for practical applications and innovations.
Locations
- Beijing, China
Salary
Estimated Salary Rangemedium confidence
300,000 - 800,000 CNY / yearly
Source: ai estimated
* This is an estimated range based on market data and may vary based on experience and qualifications.
Skills Required
- Expertise in designing and optimizing distributed training frameworks for LLMs.intermediate
- Proficiency in performance tuning and optimization of frameworks like PyTorch or VERL.intermediate
- Knowledge of high-availability architectures and stability measures for large-scale training.intermediate
- Strong collaboration skills with algorithm and platform teams.intermediate
- Experience in pushing AI large model technologies for practical applications and innovations.intermediate
Responsibilities
- Design and optimize LLM training frameworks to support high-performance training, collaborating closely with algorithm and platform teams to ensure stable, high-performance, and scalable LLM infrastructure.
- Design and build distributed training frameworks in collaboration with the platform team, supporting training of billion-parameter large models.
- Perform performance tuning and cost optimization for large model training tasks, optimizing distributed strategies in frameworks like PyTorch/VERL to improve training efficiency.
- Ensure stability and reliability by designing high-availability architectures, addressing risks such as training interruptions and data loss for long-cycle training tasks.
- Collaborate with the algorithm team to understand model requirements, provide technical advice on infrastructure, and drive customized development of open-source tools to fit business scenarios.
Tags & Categories
Answer 10 quick questions to check your fit for 元宝-大模型训练工程师 @ Tencent.
10 Questions
~2 Minutes
Instant Score
Related Books and Jobs
元宝-大模型训练工程师
Tencent
元宝-大模型训练工程师
Job Description
元宝-大模型训练工程师
📋 Job Overview
The Yuanbao Large Model Training Engineer role at Tencent involves designing and optimizing LLM training frameworks to support high-performance training of large-scale models. The position requires close collaboration with algorithm and platform teams to ensure stable, high-performance, and scalable LLM infrastructure. Key focuses include system design, performance tuning, stability assurance, and driving AI large model innovations for business applications.
📍 Location: Beijing, China
🏢 Business Unit: CSIG
📄 Full Description
1.负责设计和优化LLM训练框架，支撑LLM高性能训练。与算法团队、平台团队紧密协作，确保LLM infra的稳定、高性能、可扩展，推动AI大模型技术的落地与创新：；
2.负责设计和优化LLM训练框架，支撑LLM高性能训练。与算法团队、平台团队紧密协作，确保LLM infra的稳定、高性能、可扩展；
3.系统设计与优化：设计并搭建分布式训练框架，跟平台协作，支持千亿级参数大模型的训练；
4.性能调优与成本优化：针对大模型训练任务，优化框架（如PyTorch/VERL等）的分布式策略，提升训练效率；
5.稳定性与可靠性保障：设计高可用架构，解决训练中断、数据丢失等风险，确保长周期训练任务的稳定性；
6.协作与落地：与算法团队紧密合作，理解模型需求，提供基础设施层面的技术建议；推动开源工具的定制化开发，适配业务场景。
2.负责设计和优化LLM训练框架，支撑LLM高性能训练。与算法团队、平台团队紧密协作，确保LLM infra的稳定、高性能、可扩展；
3.系统设计与优化：设计并搭建分布式训练框架，跟平台协作，支持千亿级参数大模型的训练；
4.性能调优与成本优化：针对大模型训练任务，优化框架（如PyTorch/VERL等）的分布式策略，提升训练效率；
5.稳定性与可靠性保障：设计高可用架构，解决训练中断、数据丢失等风险，确保长周期训练任务的稳定性；
6.协作与落地：与算法团队紧密合作，理解模型需求，提供基础设施层面的技术建议；推动开源工具的定制化开发，适配业务场景。
🎯 Key Responsibilities
- Design and optimize LLM training frameworks to support high-performance training, collaborating closely with algorithm and platform teams to ensure stable, high-performance, and scalable LLM infrastructure.
- Design and build distributed training frameworks in collaboration with the platform team, supporting training of billion-parameter large models.
- Perform performance tuning and cost optimization for large model training tasks, optimizing distributed strategies in frameworks like PyTorch/VERL to improve training efficiency.
- Ensure stability and reliability by designing high-availability architectures, addressing risks such as training interruptions and data loss for long-cycle training tasks.
- Collaborate with the algorithm team to understand model requirements, provide technical advice on infrastructure, and drive customized development of open-source tools to fit business scenarios.
🛠️ Required Skills
- Expertise in designing and optimizing distributed training frameworks for LLMs.
- Proficiency in performance tuning and optimization of frameworks like PyTorch or VERL.
- Knowledge of high-availability architectures and stability measures for large-scale training.
- Strong collaboration skills with algorithm and platform teams.
- Experience in pushing AI large model technologies for practical applications and innovations.
Locations
- Beijing, China
Salary
Estimated Salary Rangemedium confidence
300,000 - 800,000 CNY / yearly
Source: ai estimated
* This is an estimated range based on market data and may vary based on experience and qualifications.
Skills Required
- Expertise in designing and optimizing distributed training frameworks for LLMs.intermediate
- Proficiency in performance tuning and optimization of frameworks like PyTorch or VERL.intermediate
- Knowledge of high-availability architectures and stability measures for large-scale training.intermediate
- Strong collaboration skills with algorithm and platform teams.intermediate
- Experience in pushing AI large model technologies for practical applications and innovations.intermediate
Responsibilities
- Design and optimize LLM training frameworks to support high-performance training, collaborating closely with algorithm and platform teams to ensure stable, high-performance, and scalable LLM infrastructure.
- Design and build distributed training frameworks in collaboration with the platform team, supporting training of billion-parameter large models.
- Perform performance tuning and cost optimization for large model training tasks, optimizing distributed strategies in frameworks like PyTorch/VERL to improve training efficiency.
- Ensure stability and reliability by designing high-availability architectures, addressing risks such as training interruptions and data loss for long-cycle training tasks.
- Collaborate with the algorithm team to understand model requirements, provide technical advice on infrastructure, and drive customized development of open-source tools to fit business scenarios.
Tags & Categories
Answer 10 quick questions to check your fit for 元宝-大模型训练工程师 @ Tencent.
10 Questions
~2 Minutes
Instant Score
Related Books and Jobs