I develop multimodal AI and large-scale deep learning systems to advance cancer diagnostics, clinical decision support, and safety-critical medical AI. My work focuses on combining foundation models, multimodal data integration, and reasoning-based AI systems to improve real-world clinical workflows.

Key Contributions

• Designed multimodal deep learning models integrating histopathology imaging, DNA methylation, gene expression, and clinical data to improve CNS tumor classification and characterize spatial tumor heterogeneity

• Developed transformer-based vision models and multiple-instance learning frameworks achieving state-of-the-art performance in multi-institutional tumor classification benchmarks

• Built agentic clinical reasoning systems using large language models and retrieval-augmented generation (RAG) to support Emergency Severity Index (ESI) triage, reducing safety-critical under-triage events while improving diagnostic accuracy

• Architected scalable ML infrastructure including distributed training pipelines, large-scale evaluation frameworks, and automated benchmarking workflows across HPC and cloud environments

• Designed production-oriented deployment pipelines with containerization, CI/CD automation, and real-time inference optimization for clinical decision-support applications

Computer Networks

Principle of Programming Languages

Compiler Writing

Operating Systems

Sensor Networks

Python

Concentration: Deep Learning, Data analytics, High performance computing

Committee member: Jun Zheng, Subhasish Mazumdar, Ramyaa, Sihua Shao

Membership: NSF EPSCoR Member, HCCS lab, NMIMT

Ph.D. Dissertation: Deep Learning Methods for Smart Meter Data Analytics and Applications

Developed and taught a 6-weeks Python course for high-school students

Involved in a multi-robot logistic system for warehouse management.

Solved traffic deadlock problems caused by maps design and traffic control.

Drove software development across multiple teams.