Current Vacancies

Research Intern – Operating Room Scene Understanding

手术室场景理解实习生

岗位背景

我们致力于推动手术室智能化，通过研究基于多视角数据的场景图（Scene Graph）建模与语义推理技术，提升医疗场景的自动分析能力。现面向对计算机视觉、多模态学习、医疗人工智能等方向有兴趣的同学，招聘实习生加入我们的科研团队。

岗位职责

1.复现手术场景图相关领域的SOTA方法

• 调研并实现手术场景图生成领域的主流和最新算法；
• 对比分析不同方法的性能，整理实验结果、撰写技术报告。

2.探索基于场景图的手术流程识别与预测

• 初步实践将场景图方法应用于手术流程识别、手术流程预测，以及手术不良事件识别等任务；
• 协助团队设计实验，分析结果并提出改进建议。

任职要求

1.计算机科学、人工智能、电子工程等相关专业本科/硕士在读；

2.具备计算机视觉或深度学习基础，熟悉主流AI开发框架（如 PyTorch、Huggingface Transformers 等）；

3.有良好的编程能力，熟练掌握 Python；

4.对多模态学习、场景图建模或医疗AI有兴趣者优先；

5.具备良好的学习能力、沟通能力与团队协作精神； 

6.每周能保证至少3天实习时间，实习期不少于3个月。

优先考虑条件

1.有 SOTA 算法复现、论文精读经验；

2.熟悉场景图（Scene Graph）、流程识别、手术视频分析等方向的相关项目或论文；  

3.有医疗影像、医疗视频相关数据处理经验；  

4.熟悉 CLIP、BLIP、LLaVA 等多模态模型的开发与应用。

申请方式

请将个人简历发送至hr02@cair-cas.org.hk。邮件主题请注明应聘[岗位名称]-[姓名]-[官网投递]。 

Background

We are committed to advancing the intelligence of the operating room through research on scene graph modeling and semantic reasoning based on multi-view data. Our goal is to enhance the automatic semantic analysis of clinical scenarios. We are looking for passionate students interested in computer vision, multimodal learning, or medical AI to join our research team as interns.

Job Responsibilities

1.Reproduce State-of-the-Art Methods in Surgical Scene Graphs 

• Conduct literature review and implement mainstream and recent algorithms in the field of surgical scene graph generation; 
• Benchmark different approaches, organize experimental results, and assist in technical report writing. 

2.Explore Scene Graph-based Surgical Workflow Recognition and Prediction

• Conduct preliminary research on applying scene graph methods to surgical workflow recognition, workflow prediction, and surgical adverse event identification;  
• Assist the team with experiment design, result analysis, and propose improvements.

 Job Requirements

1.Bachelor’s or Master’s student in Computer Science, Artificial Intelligence, Electronic Engineering, or related fields;

2.Strong foundation in computer vision or deep learning, and proficiency in mainstream AI frameworks (e.g., PyTorch, Huggingface Transformers); 

3.Solid programming skills, especially in Python;

4.Interest in multimodal learning, scene graph modeling, or medical AI is a plus;

5.Good communication skills, strong willingness to learn, and a collaborative spirit; 

6.Able to commit at least 3 days per week for a minimum of 3 months.

Preferred Qualifications

1.Experience in reproducing SOTA algorithms or reading academic papers;

2.Familiarity with scene graph, workflow recognition, or surgical video analysis projects/papers; 

3.Experience with medical image or video data processing;

4.Knowledge of multimodal models such as CLIP, BLIP, LLaVA, etc.

Application Method

Please send your resume to hr02@cair-cas.org.hk. For the email subject line, please indicate: Application for [Position Name] - [Name] - [Applied via CAIR Official Website].