My work sits at the intersection of Artificial Intelligence and Biomedicine. I specialize in deploying Large Language Models (LLMs) for clinical workflows and utilizing Foundation Models for genomic sequence analysis.

• Genomic AI: Zero-shot inference with Evo2-40B (StripedHyena architecture).
• Clinical NLP: Automating patient recruitment with LLM agents.
• Medical Imaging: 3D Point Cloud reconstruction & Bayesian Optimization.

Domain	Toolkit
LLMs	Transformers LLMs Prompt Engineering Agent
Deep Learning	DNN Gaussian Processes Bayesian Optimization
Data Eng & HPC	SQL ETL Pipelines
Languages

• Zero-shot Inference: Deployed NVIDIA's Evo2-40B (StripedHyena 2 architecture) for unsupervised recognition of bacterial sequence features.
• HPC Optimization: Built a large-scale parallel inference pipeline on High-Performance Computing (HPC) clusters.
• Innovation: Implemented Windowed Inference and temperature calibration to solve instability issues in long-sequence reasoning, significantly improving robustness.

• Patient Recruitment Agent: Led the development of an LLM-based system to extract structured data from unstructured clinical texts.
• Performance: Designed advanced Prompt Engineering strategies and semantic matching algorithms, boosting patient-trial matching efficiency by 30%.

• Algorithm Design: Developed a 3D point cloud reconstruction framework using Gaussian Processes to replace manual parameter tuning.
• Results: Reduced geometric reconstruction error to 2.8% and improved tuning efficiency by 50% through a custom multi-objective loss function.

Parexel (Clinical Research Organization) | AI/LLM Application Intern

• Focused on automating clinical trial operations using Generative AI.
• Optimized rule-based and semantic matching algorithms for patient screening.

SAS Institute | Data Analytics Intern

• Built high-dimensional Neural Network (DNN) models (AUC 0.86).
• Developed automated ETL pipelines processing 100k+ records using Python/SQL.