Designed and implemented a high-speed document aggregation pipeline in Django using Celery for parallelism, retrieving and de-duplicating ∼5,000 patient records from Carequality, CommonWell, and QHINs within minutes.
•
Developed custom CCDA document parsers compliant with HIPAA and HL7 standards to extract critical clinical data, enabling seamless integration into large language models (LLMs) for advanced insights.
•
Engineered a secure, HIPAA-compliant EHR insight solution using Next.js 14, offering healthcare providers access to patient records, vitals, chronic conditions, and address HEDIS measures via an intuitive dashboard.
•
Built a Chrome extension for seamless EHR integration, allowing healthcare personnel to access the web-based solution directly within third-party EHR applications, enhancing usability without workflow disruption.
•
Deployed the end-to-end solution on Azure Cloud, utilizing Azure Kubernetes Services (AKS), PostgreSQL, Azure Storage, Azure Container Instances, and Cloudflare to ensure scalability, performance, and security.
•
Optimized system security by integrating HIPAA-compliant practices, including secure data storage, encrypted communication with TLS authentication, along with DDoS mitigation.
Leading the development of an AI-driven ECD mobile application using Flutter for both Android and iOS platforms. Integrating large multimodal models like LLaVA using Langchain, for language and cognitive evaluation of children. Incorporated secure authentication and user management with Firebase Authentication and Firestore Database.
· As a Graduate Employee Adjunct, I served as the Section Leader for DS-UA 301: Advanced Data Science course at NYU Center for Data Science.
· Provided guidance to students in the utilization of deep learning frameworks like TensorFlow and PyTorch through hands-on lab sessions, ensuring their comprehension of these tools.
