UNIV - Staff Scientist I - Department of PHS

Medical University of South Carolina•Charleston, SC

About The Position

Designs, develops, and maintains large language model (LLM), retrieval-augmented generation (RAG), and natural language processing (NLP) systems for research across MUSC. This includes infrastructure for the Cancer Integrated Data-Enabled Resource (CIDER), which supports research at Hollings Cancer Center. The role collaborates with clinical, educational, and research teams to identify requirements and obtain data needed to build, validate, and deploy AI-enabled research tools, and partners with investigators and project teams to define requirements for new research initiatives. The Staff Scientist designs and conducts experiments to assess model accuracy, reliability, and suitability for research, and prepares manuscripts, reports, presentations, and proposal materials for scientific and institutional audiences. The role also develops, implements, and oversees procedures for secure, available, and compliant operation of AI applications and supporting infrastructure. It contributes to long-term plans for AI-enabled research tools and responds to research teams’ requests for AI-enabled tools and analyses. Duties include conducting research and engineering for CIDER; building and maintaining LLM and RAG systems for clinical data; developing retrieval, prompting, and evaluation methods; deploying AI-assisted pipelines and applications in secure environments; managing cloud-hosted LLM services, containerization, monitoring, reliability, and costs; evaluating AI methods through benchmarks and model comparisons; and collaborating with team members and project managers to complete work accurately and on schedule.

Requirements

Computer literacy.
Ability to obtain and maintain a valid driver’s license.
Ability to work rotating shifts as required.
Ability to work overtime as required.
Ability to learn and use new processes, tools, and equipment as required.
Ability to meet the listed physical requirements, including lifting and carrying up to 30 pounds unassisted and pushing or pulling up to 50 pounds unassisted.
Ability to be physically qualified by medical personnel for respirator use, initially and annually.

Responsibilities

Conduct research and engineering for CIDER in support of cancer research, including designing, building, evaluating, and maintaining LLM and RAG systems for clinical data.
Develop retrieval, prompting, and evaluation methods and validate system outputs for research projects.
Develop, deploy, and maintain AI-assisted research pipelines and applications, including LLM-assisted coding of clinical transcripts and adaptive learning tools, in secure, controlled environments.
Manage cloud-hosted LLM services, containerization, monitoring, reliability, and costs.
Design and conduct experiments to evaluate AI methods, including benchmark evaluations and model comparisons; analyze and interpret results.
Prepare manuscripts for peer-reviewed journals and conferences, as well as reports, presentations, manuscript methods sections, and proposal materials.
Gather requirements for new and ongoing funded studies by working with investigators, clinical and educational teams, and data owners to define use cases, data sources, and validation criteria.
Develop, implement, and oversee procedures to support the security, availability, and compliant operation of AI applications and supporting infrastructure.
Contribute to long-term plans for AI-enabled research tools and respond to research teams’ requests for AI-enabled tools and analyses.
Collaborate with team members and project managers through meetings and collaborative programming to complete assigned tasks accurately and on schedule.