Research

Research Interests

My research focuses on developing AI/ML solutions for oncology, with emphasis on multimodal learning, digital pathology, and clinical applications of large language models.

Current Research Projects

Multimodal Cancer Outcome Prediction

Developing deep learning frameworks like SeNMo and PARADIGM that integrate multi-omics data (genomics, transcriptomics, proteomics, epigenomics), pathology images, and clinical records for improved survival prediction across 32 cancer types.

  • Self-normalizing neural networks for pan-cancer prognostication
  • Graph-based models leveraging heterogeneous data embeddings
  • Foundation model-driven multimodal AI with HONeYBEE framework

Digital Pathology & Foundation Models

Revolutionizing diagnostic pathology through generative AI and foundation models, enabling automated tissue detection, tumor classification, and biomarker identification from histopathological images.

  • Uncertainty-aware AI for enhanced tissue detection accuracy
  • Expert evaluation of LLM reasoning in diagnostic pathology
  • Integration of vision-language models in healthcare

Early Cancer Detection

Building personalized risk assessment models using low-dose CT screening data to improve early detection of lung and breast cancer, enabling better treatment outcomes and quality of life.

  • Longitudinal fusion of LDCT data for lung cancer risk prediction
  • Post-hoc analysis of National Lung Cancer Screening Trial
  • Personalized future cancer risk assessment

Medical Large Language Models

Evaluating and deploying private, local LLMs for medical data processing, including extraction of discrete variables from clinical records, ICD billing, and cancer registry automation.

  • M-Llama: Local LLM for cancer registry data extraction
  • Consensus-based reasoning for pathology report analysis
  • Diagnostic reasoning evaluation in radiology and pathology

Privacy-Preserving Federated Learning

Implementing federated learning approaches for multi-institutional medical imaging research, enabling collaborative model training while preserving patient privacy and data security.

  • Uncertainty quantification in medical imaging
  • Distributed intelligence for early lung cancer detection
  • Foundation federated learning for multi-source data integration

Immunotherapy Response Prediction

Developing accurate biomarkers to identify patient sub-populations that will benefit most from immunotherapy treatment, improving treatment efficacy and resource allocation.

  • Genetic markers and treatment response analysis
  • Survival outcome prediction for immunotherapy
  • Targeted prevention strategies development

Selected Presentations

Oral Presentation • 2025

Beyond Black-Box AI: Assessment of Pathology-Specific Reasoning in Generative Models by Pathologists

Pathology Visions 2025

San Diego, California

Poster Presentation • 2025

Enhancing Tissue Detection Accuracy in Digital Pathology through Uncertainty-Aware AI

Pathology Visions 2025

San Diego, California

Abstract • 2025

Using Patient Embeddings From Foundation Models for Enhanced Survival Analysis in Lung Squamous Cell Carcinoma

American Thoracic Society (ATS) 2025

Abstract • 2025

PARADIGM: An Embeddings-based Multimodal Learning Framework with Foundation Models and Graph Neural Networks

AACR Annual Meeting 2025

Poster Presentation • 2025

Harnessing Distributed Intelligence with Privacy Preservation: Federated Learning for the Early Detection of Lung Cancer

NVIDIA GTC 2025

San Jose, USA

Presentation • 2025

AI-Driven Extraction of Key Clinical Data from Pathology Reports to Enhance Cancer Registries

USCAP 114th Annual Meeting

Boston, Massachusetts

Oral Presentation • 2024

Extraction of Discrete Information from Pathology Reports Using Local and Private LLMs

Digital Pathology Association, Pathology Visions 2024

Oral Presentation • 2024

Harnessing Distributed Intelligence with Privacy Preservation: Federated Learning in Oncology

Cancer Biomarkers AI and Bioinformatics Workshop 2024

Honors & Awards

1st Place - Bio-Data Club Hackathon 2024

Moffitt Cancer Center

$25,000 pilot grant - M-Llama: Moffitt's Local LLM for Cancer Registry Data Extraction

3rd Place - Bio-Data Club Hackathon 2023

Moffitt Cancer Center

Generating, visualizing, and quantitatively analyzing graphs of multi-omics data

3MT (Three Minute Thesis) Competition Selection

University of South Florida

Selected for the university-wide competition

Summa Cum Laude

University of South Florida

Ph.D. in Electrical Engineering with GPA 4.0/4.0, Dean's List

Graduate Assistantship Award

University of South Florida

Full funding for doctoral studies

Graduate Assistantship Award

Rowan University

Full funding for doctoral studies

Tau Beta Pi Member

Engineering Honor Society

Membership for being an honor engineering student

Federal Board Scholarship for Higher Education

Government of Pakistan (FBISE)

Merit-based scholarship for academic excellence

Professional Memberships

American Association for Cancer Research (AACR)

2024 - Present

American Thoracic Society (ATS)

2024 - Present

Digital Pathology Association (DPA)

2023 - Present

Tau Beta Pi Engineering Honor Society

2023 - Present

Institute of Electrical and Electronics Engineers (IEEE)

2021 - 2024