• Led the integration of deep learning models with climate data, significantly improving model performance by leveraging large-scale datasets and applying advanced data preprocessing and feature engineering techniques.• Designed and implemented multitask learning frameworks and meta-learning approaches for climate teleconnection predictions, optimizing the model’s scalability and efficiency for complex spatio-temporal data.• Adjusted Vision-Language Models (VLMs) to recognize climate-related extreme events by embedding climatological information. Explored the model’s potential predictive capabilities, enabling it to respond to image-based queries by providing possible future climate trends and insights based on the input data.

• Integrated machine learning with physics-based models to accelerate scientific discovery, particularly in climate-related predictions, improving forecast accuracy for extreme weather events.• Applied implicit neural representations to enable the transformation of climate data at arbitrary spatial resolutions, taking into account the inherent spatial and temporal properties of climate datasets.• Employed knowledge distillation techniques to facilitate information transfer between physics-based models and data-driven models, improving the synergy between mechanistic understanding and data-based predictions for more accurate and efficient forecasting.

• Led the management and enhancement of the NIDILRR Innovation Management System (NIMS) database using MongoDB, encompassing approximately 2,000 technology-related grants. Utilized AWS cloud architecture to handle large-scale data processing and analysis, ensuring the system’s scalability and reliability for continuous dataset expansion; • Provided periodic descriptive analysis of the Assistive Technology Tech Transfer (ATTT) dataset, identifying trends and patterns in rehabilitation technology. Implemented data mining techniques to uncover insights that informed decision-making and improved innovation management processes.• Designed and implemented an AI agent using large language models (LLMs) to evaluate the potential outcomes and commercial product, which enhanced the ability to assess grant impact and prioritize funding decisions.

• Spearheaded the development, customization, and maintenance of the ISWP website, utilizing WordPress and PHP to enhance site functionality and user experience. Implemented innovative features and improvements that significantly increased user engagement and accessibility;• Led data mining and analysis initiatives based on user interactions and feedback from across the globe collected via the website. Applied visualization to translate complex datasets into actionable insights, driving strategic decisions that improved service delivery and outreach.