- Implemented a data architecture for chatbot ingestion into AWS RDS using ETL processes.- Developed and deployed a chatbot using WhatsApp and OpenAI APIs (Django backend) on Heroku, achieving over 2000 users in LATAM for oncology patient support.- Created dashboards in DataLooker to track company health metrics such as cohort churn, daily, weekly, monthly activity and RFM segmentation. This allowed the company to make decisions for experiments.- Implemented RAG feature to chatbot using pinecone database, enhanced customer engagement with chatbot by 30%

Developed end to end health claim models:- Classification model to predict in the following 3 months outpatient claims. Allowed redirection of patients to virtual appointments reducing costs (75-100k $ yearly)- Regression model to predict budget ceiling of clinics. Allowed insurer to make better decisions on which letters of guarantee to accept, preventing overcosts.

Collaborated with a team of 5 engineers to:- Design (CAD) and implement a health triage machine.- Develop a database using SQL and Python to ingest data from the machine efficiently.- Lead an experimental pilot study with 100+ participants, validating the machine and reducing health evaluation times by one-third.