LLM Research/Engineering Intern

Job Type: Full-time Intern Location: Beijing/Shanghai  Group Introduction: The Microsoft DKI team is at the forefront of exploring cutting-edge research and applications of multimodal large models. We have developed multiple open-source projects such as Wizard, TaskWeaver, and UFO. Our work focuses on integrating innovative LLM technologies with cloud systems and data to serve various applications like automated fault handling, intelligent log analysis, Text2SQL, and code intelligence.  Job Responsibilities (Choose between Engineering or Research):  Engineering Intern: 
  • Participate in the design and engineering development of next-generation Agent frameworks, including LLM Agent and GUI Agent. 
  • Engage in the collection, analysis, and processing of multimodal data to support Agent-specific model reasoning and training, and optimize the training pipeline. 
  • Facilitate the integration and application of the Agent framework in Microsoft products. 
Research Intern: 
  • Fine-tune and deploy multimodal large models, involving technologies such as CoT, Agent, SFT, RLHF, DeepSpeed, and vllm. 
  • Contribute to prompt design, optimization, and compression for multimodal large models. 
  • Publish research findings in reputable journals or top academic conferences. 
Qualifications: 
  • Proficient in training deep learning models with substantial hands-on project or deployment experience. 
  • Some project experience related to large language models (LLM). 
  • Prior experience with Agent-related projects is preferred. 
  • Strong engineering skills, diligent and responsible, with excellent communication abilities. 
  • Publications in top academic conferences or journals are an advantage. 
  • Must be available for a full-time internship for 6 months or more. 
Internship Duration Requirements: Must obtain permission from your academic advisor and commit to at least 6 months of internship.  Please send your English or Chinese resume (in PDF/Word format) to: intern_hiring@microsoft.com. Please include “LLM Research/Engineering Intern” in the email subject line.    岗位名称：大模型研究/工程实习生 工作性质：全职实习生 工作地点：北京/上海  团队介绍： 我们是微软DKI团队，致力于多模态大模型的前沿研究与应用，开发了Wizard、TaskWeaver和UFO等多个开源项目。我们的工作重点是将大模型创新技术与云系统及数据有机结合，服务于自动故障处理、智能日志分析、Text2SQL、代码智能等应用场景。  岗位职责（方向可选）：  工程实习生： 
  • 参与下一代Agent框架的设计和工程开发，包括LLM Agent、GUI Agent等。 
  • 参与多模态数据的收集、分析和处理，以支持Agent专属模型推理和训练，并优化训练流程。 
  • 推动Agent框架在微软产品中的集成和应用落地。 
研究实习生： 
  • 负责多模态大模型的微调和部署，涉及CoT、Agent、SFT、RLHF、DeepSpeed、vllm等技术。
  • 参与多模态大模型的prompt设计、优化和压缩。 
  • 将研究成果发表在权威期刊或顶级学术会议上。 
岗位要求： 
  • 熟练掌握深度学习模型的训练，具备丰富的深度学习项目或实际落地经验。 
  • 有一定大语言模型（LLM）相关项目经验。 
  • 具备Agent等相关项目经验者优先。 
  • 具备优秀的工程动手能力，工作踏实认真，沟通能力强。 
  • 在顶级学术会议或期刊发表过论文者优先。 
  • 每周全职工作5天，能够实习6个月以上 
实习时间要求： 需获得导师许可，并保证至少6个月的实习周期。  请将完整的中/英文简历（PDF/Word形式）发送至：intern_hiring@microsoft.com。邮件标题中请注明“大模型研究/工程实习生”。   

Microsoft is an equal opportunity employer. All qualified applicants will receive consideration for employment without regard to age, ancestry, citizenship, color, family or medical care leave, gender identity or expression, genetic information, immigration status, marital status, medical condition, national origin, physical or mental disability, political affiliation, protected veteran or military status, race, ethnicity, religion, sex (including pregnancy), sexual orientation, or any other characteristic protected by applicable local laws, regulations and ordinances. If you need assistance and/or a reasonable accommodation due to a disability during the application process, read more about requesting accommodations.