锦鲤求职岗位描述
岗位职责
Introduction
This subtrack is hosted by the STOmics Pathology team at BGI Research, Chongqing. Using BGI's spatial multi-omics platforms, we integrate high-resolution tumor histopathology with spatial transcriptomics, single-cell spatial profiling, spatial proteomics, and clinical data to connect tissue morphology with molecular states and support translational cancer research. Our research focuses on the following areas.
Pathology foundation models. Develop and adapt SURF, our pathology vision foundation model, and STOP, our multimodal pathology foundation model, to learn fine-grained and cross-modal representations for analysis at the cell, tissue patch, whole-slide image (WSI), and patient levels.
Clinical prediction models. Build and validate models for cell and tissue analysis, molecular biomarker prediction, cancer risk assessment, and treatment response prediction.
Pathology tools and AI agents. Develop WSI retrieval and multitask analysis workflows for research and clinical applications. Model and workflow deployment on DCS Cloud.
Biological discovery. Integrate pathology and spatial multi-omics to characterize tumor heterogeneity and spatial cellular interactions, identify candidate therapeutic targets, and investigate mechanisms associated with disease progression and treatment response.
Affiliated Branch
Southwest Branch (Chongqing)
岗位要求
(1)Preferred Background
Computer Science, Artificial Intelligence, Bioinformatics, Biomedical Engineering, Applied Mathematics, Statistics, or related disciplines. Applicants from biological or medical sciences are also welcome if they demonstrate strong quantitative and programming skills. Prior training in pathology or spatial omics is advantageous but not required.
(2)Core Technical Skills
Proficiency in Python and practical experience with PyTorch or an equivalent deep learning framework. A solid foundation in machine learning, linear algebra, probability, and statistics; ability to perform data preprocessing, quality control, and model evaluation. Familiarity with Linux, Git, and reproducible computational workflows.
(3)Modeling Competency
Ability to translate biological or clinical questions into computational tasks, implement baseline models, and adapt pretrained models through feature extraction, fine-tuning, or multimodal fusion. Familiarity with at least one relevant approach, such as image classification, segmentation, multiple instance learning, or representation learning. Understanding of sound experimental design, patient-level data splitting, data leakage prevention, and assessment of model generalization.
(4)Preferred Qualifications
Research or project experience in computational pathology, medical image analysis, spatial or single-cell omics, or multimodal learning. Experience with WSI processing, self-supervised or weakly supervised learning, survival analysis, external validation, or model deployment is a plus. Publications, open-source contributions, or well-documented research projects are valued.
(5)Personal Traits
Scientific curiosity, independent thinking, and a willingness to learn across AI, biology, and medicine. Careful experimental practice, perseverance with noisy real-world data, and respect for research integrity and data confidentiality. Effective teamwork and English communication skills for reading papers, discussing results, and writing scientific reports.