Location: Houston, TX
Cap-Exempt H-1B Position — No lottery required
## **Summary** The Postdoctoral Associate will develop next-generation AI models for large-scale perturbation modeling in brain tumors. The project will involve building and applying state-of-the-art machine learning approaches, including foundation models, variational autoencoders (VAEs), and transformer-based architectures, to integrate single-cell and multi-omic datasets. The goal is to decode tumor cellular heterogeneity and tumor microenvironment interactions, and to identify targetable genes, pathways, and therapeutic strategies at single-cell resolution. Baylor College of Medicine typically follows similar to the NIH stipulated stipend guidelines for Postdoctoral Associates. ## **Job Duties** - Develops and implements AI models for perturbation prediction: - Designs, trains, and evaluates machine learning models (e.g., transformer-based architectures, VAEs, and foundation models) to predict cellular responses to genetic and pharmacologic perturbations. This includes preprocessing large-scale single-cell and multi-omic datasets, defining model architectures, optimizing training pipelines on GPU clusters, and benchmarking against existing methods. - Integrate and analyze large-scale single-cell and multi-omic: - Processes and harmonizes scRNA-seq, scATAC-seq, and related datasets across brain tumor cohorts. - Performs downstream analyses such as cell state annotation, pathway enrichment, and tumor–tumor microenvironment interaction modeling to generate biologically meaningful insights. - Leads computational research projects and method development. - Performs other job-related duties as assigned. ## **Minimum Qualifications** - MD or Ph.D. in Basic Science, Health Science, or a related field. - No experience required. ## **Preferred Qualifications** - Ph.D. in Computational Biology, Bioinformatics, Computer Science or a related quantitative field. - Strong background in machine learning and statistical modeling, with experience in deep learning frameworks (e.g.