Postdoctoral Research Fellowship in Machine Learning
Wang Laboratory, University of Michigan, Ann Arbor, MI, United States
Employer Description:
PI: Thomas D Wang, MD,PhD, University of Michigan
Project goal is to develop machine learning algorithms for a robotically-assisted imaging platform to enhance visualization of tumor margins during intra-operative surgery.
Post-doctoral research opportunities are available in the Wang Lab at the University of Michigan to join a highly collaborative, multi-disciplinary team to develop an end-to-end solution for a multi-modal imaging system to visualize tumor margins during intra-operative surgery. A confocal photoacoustic endomicroscope will be developed to detect next generation contrast agents for cancer specific targets. The imaging system will be verified first in pre-clinical models followed by in human subjects. Clinical studies will be performed by a highly experienced international team of surgeons in collaboration with a global leader in robotic-assisted minimally invasive surgery. The integrated solution will improve precision for tumor localization and differentiation from normal tissues to improve patient outcomes. The successful applicant will have an opportunity to develop machine learning algorithms for real time clinical interpretation of in vivo fluorescence and photoacoustic images. A portable platform will be used for fast processing speeds that will enable rapid clinical decision making in the operating room. Training of deep learning models will be performed using datasets of images collected in vivo from human subjects.
Requirements: Candidates with a Ph.D. in computer science, engineering (e.g. electrical, biomedical, mechanical, aeronautics, etc), bioinformatics, statistics, or a related discipline with project experience in machine learning, deep learning, and natural language processing are invited to apply. Programming experience with MATLAB, Python, and C++ is desirable. International investigators are welcome. Relevant skills include the following:
· Expertise in relevant software programming languages
· Experience with UNet architecture with MobileNet V2 backbone
· Understanding of ResNet, Inception, and Xception.
· Familiarity with Classification and Regression Tree (CART) and Random Forest (RF) methods
· Proficiency in image denoising, motion artifact mitigation, and feature segmentation.
· Knowledge of robotics systems, control systems, and systems engineering.
· Excellent organizational and communication skills.
Research environment: The University of Michigan is committed to diversity and equality in education and employment, and is an ideal training ground for postdocs who want to further develop their skills in pursuit of a career in academia or industry. Salary will be commensurate with the applicant’s experience and in compliance with NIH guidelines.
The Wang Laboratory is located in the Biomedical Science Research Building (BSRB) on the University of Michigan medical campus, and is located next to the University Hospital and Rogel Comprehensive Cancer Center. Lab resources for use in this project include Dell Precision 5820 tower systems with Intel Xeon Processor W-2295 (18 cores, 3.0 GHz), 128GB DDR4 RAM, 1/16 TB of internal/external disk space, and Linux operating system. The Advanced Research Computing (ARC) facility provides computing resources, including high performance computing, machine learning, storage, data, and private/public cloud analytical resources. ARC offers a professional dedicated IT staff to monitor and maintain all systems for security and stability as well as provide end-user support, training, and outreach. The "Great Lakes" supercomputer, a Linus cluster of over 16,000 cores, a large memory of over 1.5 TB/node, and GPU accelerators is available for use. This resource provides a full application and software development library of over 100 titles of both open source and commercial tools. For storage, ARC provides "Turbo", an HHD array of high speed storage of over 10PB of capacity. In addition, "Data Den" can be used for massive-capacity and long term data archive as a tape-based storage system that provides over 20PB of replicated capacity. The Michigan Institute for Data Sciences (MIDAS) provides resources to develop generative artificial intelligence models, including UM GPT, Maizey, and GPT Toolkit. These methodologies can be used for general purpose research, natural language text, and custom training datasets.
Candidates should submit CV and contact information by email:
Thomas D. Wang, M.D.,Ph.D.
Professor of Medicine, Biomedical Engineering, and Mechanical Engineering
H. Marvin Pollard Collegiate Professor of Endoscopy Research
University of Michigan
109 Zina Pitcher Pl. BSRB 1522
Ann Arbor, MI 48109-2200
Email: tdwangmd@gmail.com
https://sites.google.com/a/umich.edu/wang_lab