Job Summary

The Complex Cellular Phenotype Analysis Group led by Dr. Lit-Hsin Loo builds computational methods and machine-learning models to measure and predict how cells, tissues, and patients respond to drugs, chemicals, or other perturbations. Recent work from the group spans spatial multi-omics profiling of human tumor tissues and high-throughput imaging-based cellular phenotypic profiling for drug toxicity or target predictions. The group is now extending this work toward AI-driven virtual cells that can simulate disease progression and therapeutic response in silico.

The Position

The successful candidate will develop foundation models, virtual cell models, and related deep learning approaches to predict cellular responses to drug or genetic perturbations, based on scRNA-seq, spatial transcriptomics, proteomics, and/or metabolomics data from both public atlas-scale resources and in-house data. We are open to two types of candidates: those with strong computational backgrounds who are interested in developing novel AI models or algorithms for biological applications, or those with strong interdisciplinary backgrounds who are interested in applying state-of-the-art models to solve challenging biological problems. We will also consider candidates for both Postdoctoral Fellow (with PhD degree) or Research Officer (with Master degree). The roles and responsibilities will be commensurate with the candidate’s experience and training. He/she will have the opportunity to work in a highly stimulating and collaborative environment and to translate the work into prototypes or products with real-world impacts.

Key Responsibilities

• Compile and pre-process large-scale training and test datasets.
• Design, train, and benchmark foundation and virtual cell models, including transfer to unseen compounds, perturbations, cell types, and doses.
• Implement and optimize computational algorithms, establish rigorous benchmarks, and maintain a reproducible and testable code base and models.
• Work with experimental collaborators to apply the models and design validation experiments.
• Present at international conferences and publish in high-impact journals.
• Senior candidates may also have the opportunity to plan and lead new projects.

Requirements

• PhD or Master's degree in machine learning, computer science/engineering, computational biology, bioinformatics, statistics, biology, or a related field.
• Strong Python and R programming skills, with comfort training models at scale on GPU/HPC.
• Previous experience in at least one of the following areas is required: single-cell or spatial transcriptomics analysis, multiplex tissue image analysis, generative or self-supervised learning, graph neural networks, or transformer models.
• Previous experience in any of the following areas is highly desirable: foundation model training or application, perturbation atlases and benchmarks, multi-omics data integration, or LLM-based scientific workflows.
• First-author publications or well-documented open-source contributions are required for postdoctoral fellow candidates.