Experience

My work at the Kiemen Lab focused on developing a pipeline for the segmentation of Pancreatic Epithelial Neoplasms (PanIN) from Low-Grade (LG) to High-Grade (HG) as a gradient, on top of the existing CODA codebase. This allows for the identification of PanINs that are most likely to become Pancreatic Ductal Adeno-Carcinoma (PDAC).

• Developed a DeepLab-ResNet50 segmentation model pipeline for PanIN on whole-slide pathology images (81% macro-F1, >93% recall).
• Curated and quality-controlled 80+ pathology annotations, standardizing labels across epithelial, stromal, and PanIN regions.

Worked as a full-time data scientist on applied machine learning and AI initiatives in healthcare and enterprise contexts, owning end-to-end development of production-oriented tools and proof-of-concept systems in collaboration with cross-functional stakeholders.

• Led development of a clinical pedigree charting tool in collaboration with healthcare stakeholders, achieving 70–80% operational efficiency improvements over PowerPoint-based workflows.
• Designed and presented proof-of-concept healthcare AI systems, including an LLM-based clinical chatbot using parameter-efficient fine-tuning and retrieval-augmented generation (RAG), to evaluate feasibility of cost-efficient, open-source models.

Worked as a full-time data scientist on applied machine learning and AI initiatives in healthcare and enterprise contexts, owning end-to-end development of production-oriented tools and proof-of-concept systems in collaboration with cross-functional stakeholders.

• Led development of a clinical pedigree charting tool in collaboration with healthcare stakeholders, achieving 70–80% operational efficiency improvements over PowerPoint-based workflows.
• Designed and presented proof-of-concept healthcare AI systems, including an LLM-based clinical chatbot using parameter-efficient fine-tuning and retrieval-augmented generation (RAG), to evaluate feasibility of cost-efficient, open-source models.

Conducted translational cancer biology research focused on tumor–immune interactions, investigating cellular communication mechanisms and therapeutic strategies in metastatic and breast cancer models. Contributed to experimental design, data analysis, and scientific communication within a multidisciplinary research environment.

• Investigated tumor–immune cell interactions, identifying mitochondrial transfer from macrophages to bladder cancer cells via nanotubules as a potential metabolic hijacking mechanism.
• Evaluated a therapeutic strategy for triple-negative breast cancer (TNBC), improving assay reproducibility and mRNA–protein complex presentation through protocol optimization.
• Authored and presented a first-author research poster on immune–cancer cell dynamics at the Orthopedic Research Retreat 2023.