《王者荣耀》强化学习算法工程师（游戏AI多智能体方向）
Tencent
《王者荣耀》强化学习算法工程师（游戏AI多智能体方向）
Job Description
《王者荣耀》强化学习算法工程师（游戏AI多智能体方向）
📋 Job Overview
The Reinforcement Learning Algorithm Engineer position at Tencent focuses on developing and implementing RL and imitation learning solutions in Unreal Engine 5 for King of Glory, targeting gameplay development like combat AI, automated testing, and numerical analysis. The role emphasizes innovations in multi-agent collaboration, complex map exploration, and human-like behaviors, while designing deployable strategies and efficient training schemes. Responsibilities include end-to-end collaboration from training to inference, experiment management, and staying updated with cutting-edge methods to build practical toolkits.
📍 Location: Chengdu, China
🏢 Business Unit: IEG
📄 Full Description
2.围绕多智能体协同、复杂地图探索与类人行为等方向开展算法创新与产品化落地；
3.设计可部署的策略与网络结构（如多头、注意力/图建模、层次化控制等），兼顾性能与可解释性；
4.制定高效率训练方案（任务分解、课程学习、奖励与随机化），在复杂训练目标中实现短迭代；
5.与工程协作完成端到端（训练—评测—导出—推理接入) 闭环，建立指标与A/B评估体系；
6.进行实验管理与复现，推动可视化与调参工具完善；跟进前沿并形成可落地的方法库。
🎯 Key Responsibilities
- 在虚幻5游戏环境中研发并落地强化学习/模仿学习方案, 服务于研发项目的玩法开发(战斗AI)，自动测试，数值分析等目标
- 围绕多智能体协同、复杂地图探索与类人行为等方向开展算法创新与产品化落地
- 设计可部署的策略与网络结构（如多头、注意力/图建模、层次化控制等），兼顾性能与可解释性
- 制定高效率训练方案（任务分解、课程学习、奖励与随机化），在复杂训练目标中实现短迭代
- 与工程协作完成端到端（训练—评测—导出—推理接入) 闭环，建立指标与A/B评估体系
- 进行实验管理与复现，推动可视化与调参工具完善；跟进前沿并形成可落地的方法库
🛠️ Required Skills
- 强化学习 (Reinforcement Learning)
- 模仿学习 (Imitation Learning)
- 多智能体协同 (Multi-Agent Collaboration)
- 复杂地图探索 (Complex Map Exploration)
- 类人行为 (Human-like Behaviors)
- 虚幻5游戏环境 (Unreal Engine 5)
- 网络结构设计 (Network Structures like Multi-Head, Attention/Graph Modeling, Hierarchical Control)
- 训练方案制定 (Training Schemes: Task Decomposition, Curriculum Learning, Rewards & Randomization)
- 端到端闭环 (End-to-End Pipeline: Training-Evaluation-Export-Inference)
- 实验管理与复现 (Experiment Management & Reproducibility)
- 可视化与调参工具 (Visualization & Hyperparameter Tuning Tools)
Locations
- Chengdu, China
Salary
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
- 强化学习 (Reinforcement Learning)intermediate
- 模仿学习 (Imitation Learning)intermediate
- 多智能体协同 (Multi-Agent Collaboration)intermediate
- 复杂地图探索 (Complex Map Exploration)intermediate
- 类人行为 (Human-like Behaviors)intermediate
- 虚幻5游戏环境 (Unreal Engine 5)intermediate
- 网络结构设计 (Network Structures like Multi-Head, Attention/Graph Modeling, Hierarchical Control)intermediate
- 训练方案制定 (Training Schemes: Task Decomposition, Curriculum Learning, Rewards & Randomization)intermediate
- 端到端闭环 (End-to-End Pipeline: Training-Evaluation-Export-Inference)intermediate
- 实验管理与复现 (Experiment Management & Reproducibility)intermediate
- 可视化与调参工具 (Visualization & Hyperparameter Tuning Tools)intermediate
Responsibilities
- 在虚幻5游戏环境中研发并落地强化学习/模仿学习方案, 服务于研发项目的玩法开发(战斗AI)，自动测试，数值分析等目标
- 围绕多智能体协同、复杂地图探索与类人行为等方向开展算法创新与产品化落地
- 设计可部署的策略与网络结构（如多头、注意力/图建模、层次化控制等），兼顾性能与可解释性
- 制定高效率训练方案（任务分解、课程学习、奖励与随机化），在复杂训练目标中实现短迭代
- 与工程协作完成端到端（训练—评测—导出—推理接入) 闭环，建立指标与A/B评估体系
- 进行实验管理与复现，推动可视化与调参工具完善；跟进前沿并形成可落地的方法库
Tags & Categories
Answer 10 quick questions to check your fit for 《王者荣耀》强化学习算法工程师（游戏AI多智能体方向） @ Tencent.
Related Books and Jobs
《王者荣耀》强化学习算法工程师（游戏AI多智能体方向）
Tencent
《王者荣耀》强化学习算法工程师（游戏AI多智能体方向）
Job Description
《王者荣耀》强化学习算法工程师（游戏AI多智能体方向）
📋 Job Overview
The Reinforcement Learning Algorithm Engineer position at Tencent focuses on developing and implementing RL and imitation learning solutions in Unreal Engine 5 for King of Glory, targeting gameplay development like combat AI, automated testing, and numerical analysis. The role emphasizes innovations in multi-agent collaboration, complex map exploration, and human-like behaviors, while designing deployable strategies and efficient training schemes. Responsibilities include end-to-end collaboration from training to inference, experiment management, and staying updated with cutting-edge methods to build practical toolkits.
📍 Location: Chengdu, China
🏢 Business Unit: IEG
📄 Full Description
2.围绕多智能体协同、复杂地图探索与类人行为等方向开展算法创新与产品化落地；
3.设计可部署的策略与网络结构（如多头、注意力/图建模、层次化控制等），兼顾性能与可解释性；
4.制定高效率训练方案（任务分解、课程学习、奖励与随机化），在复杂训练目标中实现短迭代；
5.与工程协作完成端到端（训练—评测—导出—推理接入) 闭环，建立指标与A/B评估体系；
6.进行实验管理与复现，推动可视化与调参工具完善；跟进前沿并形成可落地的方法库。
🎯 Key Responsibilities
- 在虚幻5游戏环境中研发并落地强化学习/模仿学习方案, 服务于研发项目的玩法开发(战斗AI)，自动测试，数值分析等目标
- 围绕多智能体协同、复杂地图探索与类人行为等方向开展算法创新与产品化落地
- 设计可部署的策略与网络结构（如多头、注意力/图建模、层次化控制等），兼顾性能与可解释性
- 制定高效率训练方案（任务分解、课程学习、奖励与随机化），在复杂训练目标中实现短迭代
- 与工程协作完成端到端（训练—评测—导出—推理接入) 闭环，建立指标与A/B评估体系
- 进行实验管理与复现，推动可视化与调参工具完善；跟进前沿并形成可落地的方法库
🛠️ Required Skills
- 强化学习 (Reinforcement Learning)
- 模仿学习 (Imitation Learning)
- 多智能体协同 (Multi-Agent Collaboration)
- 复杂地图探索 (Complex Map Exploration)
- 类人行为 (Human-like Behaviors)
- 虚幻5游戏环境 (Unreal Engine 5)
- 网络结构设计 (Network Structures like Multi-Head, Attention/Graph Modeling, Hierarchical Control)
- 训练方案制定 (Training Schemes: Task Decomposition, Curriculum Learning, Rewards & Randomization)
- 端到端闭环 (End-to-End Pipeline: Training-Evaluation-Export-Inference)
- 实验管理与复现 (Experiment Management & Reproducibility)
- 可视化与调参工具 (Visualization & Hyperparameter Tuning Tools)
Locations
- Chengdu, China
Salary
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
- 强化学习 (Reinforcement Learning)intermediate
- 模仿学习 (Imitation Learning)intermediate
- 多智能体协同 (Multi-Agent Collaboration)intermediate
- 复杂地图探索 (Complex Map Exploration)intermediate
- 类人行为 (Human-like Behaviors)intermediate
- 虚幻5游戏环境 (Unreal Engine 5)intermediate
- 网络结构设计 (Network Structures like Multi-Head, Attention/Graph Modeling, Hierarchical Control)intermediate
- 训练方案制定 (Training Schemes: Task Decomposition, Curriculum Learning, Rewards & Randomization)intermediate
- 端到端闭环 (End-to-End Pipeline: Training-Evaluation-Export-Inference)intermediate
- 实验管理与复现 (Experiment Management & Reproducibility)intermediate
- 可视化与调参工具 (Visualization & Hyperparameter Tuning Tools)intermediate
Responsibilities
- 在虚幻5游戏环境中研发并落地强化学习/模仿学习方案, 服务于研发项目的玩法开发(战斗AI)，自动测试，数值分析等目标
- 围绕多智能体协同、复杂地图探索与类人行为等方向开展算法创新与产品化落地
- 设计可部署的策略与网络结构（如多头、注意力/图建模、层次化控制等），兼顾性能与可解释性
- 制定高效率训练方案（任务分解、课程学习、奖励与随机化），在复杂训练目标中实现短迭代
- 与工程协作完成端到端（训练—评测—导出—推理接入) 闭环，建立指标与A/B评估体系
- 进行实验管理与复现，推动可视化与调参工具完善；跟进前沿并形成可落地的方法库
Tags & Categories
Answer 10 quick questions to check your fit for 《王者荣耀》强化学习算法工程师（游戏AI多智能体方向） @ Tencent.
Related Books and Jobs