SDE working on cloud computing and serverless development
San Francisco, California, United States
Designed & built internal model context protocol based AI support system handling hundreds level 2C & 2B support cases
Designed & Deployed Scalable Cloud Infrastructure: Led computing resource deployment and monitoring setup using AWS CDK, optimizing cloud resource management.
Architected Backend Schema for Longevity Health Tech Ecosystem: Developed scalable data models to support patient, clinician, and vendor interactions.
Integrated Vendor Services for Subscription & Order Processing: Onboarded multiple vendors for user subscriptions, payments, and order tracking using AWS API Gateway, Lambda, and CDK—processing thousands of orders.
Automated Order Filtering & Fulfillment: Built an automated system using AWS EventBridge Scheduler, API Gateway, and Lambda to track and process thousands of product orders.
Developed Patient & Clinician Portals: Implemented AWS Cognito-based authentication, data storage, and API-driven interactions to enhance patient-clinician connectivity.
Migrated Company Database from DynamoDB to DocumentDB: Designed and executed a smooth cloud database migration, improving query performance and scalability.
Built Webhook Event Handlers for Vendor Integrations: Automated event-driven workflows, ensuring seamless third-party service interactions.
Implemented Automated Email & SMS Notifications: Developed AWS SES and EventBridge-based messaging systems, improving user engagement.
Enabled Seamless Telehealth Appointment Scheduling: Integrated event scheduling and meeting platforms in the backend, supporting hundreds of online appointments.
Enhanced User Communication with SMS Notifications: Onboarded SMS service vendors and implemented automated notifications using AWS EventBridge and Lambda.
Designed & Built a Backend User Activity Tracking System: Developed a robust tracking system to monitor and analyze user interactions, enabling data-driven product improvements.
2023 — 2023
San Francisco Bay Area
Built MySQL and MongoDB database for a web3 Event Web App and maintained million-level data storage.
Designed and implemented multiple sets of CRUD API functions supporting Web App UI features on separate Backend servers for company official website and flagship products using Node.js and Java Spring Boot.
Implemented various Frontend features on the Business-side event planner tool to enable cross-team sync during event planning and optimized performances of existing components using Vue.js and React.js.
Tested and deployed both project Frontend Nginx server and backend Node, Spring Boot servers on AWS EC2.
Virginia, United States
Implemented a FastAPI webhook listening service to retrieve updated documents from client box account folders. Improved the integration efficiency between external document storage site and company backend NLP server by 40%.
Redesigned the backend API-gateway using FastAPI framework to assist direct document uploads from the server users. Increased the upload speed of unprocessed legal contract documents by 30%.
Organized the document processing pipeline using RabbitMQ queue. Improved the document markup time by 50%.
Tested the webhook listening service using Ngrok end-points. Created Docker image for the application package.
Boston, Massachusetts, United States
Conducted Capstone project under mentorship from Dr. Li-wei H. Lehman at MIT - Lab for Computational Physiology.
Performed data imputation with padding method to preprocess Mechanical Ventilation measurement data of ARDS patients for predictive modeling in Python.
Applying Domain Adaptation in Machine Learning and Counterfactual Prediction model trained with traditional ARDS patient data to perform treatment outcome prediction for new Covid-19 patient cohort using Keras in Python.
University of California, Los Angeles
Built a predictor for food addiction levels from early life adversity with PLS-DA and Neural Network models in Python
Employed PCA method to reduce dataset dimension and employed VIF to test multicollinearity in predictor variables
Tuned final NN model and obtained a high prediction accuracy for food addiction from Early Trauma data at 93%
Education
The Johns Hopkins University
MSE
Harvard Medical School
Master of Biomedical Informatics
UCLA