AWS Cloud ML
Risk > 0.45
Heart Health Alert System
Real-time cardiac risk prediction and alerting on AWS.
140+ patient-day records
Vitals land in S3, then Spark cleans and aggregates the weekly signal.
End-to-end architecture
S3 → EMR Spark → SageMaker → Lambda → SNS → Athena.
High-risk alert
Patient P-007 · Risk score 0.78
SNS email notification triggered
Prediction becomes action
Lambda scores records and sends alerts when the threshold is crossed.
Healthcare vitals data needs scalable preprocessing, prediction, and alerting so risk signals can be detected quickly instead of staying buried in raw records.
My Role
Cloud data pipeline designer
Architecture
S3 patient vitals to Spark on EMR to SageMaker prediction to Lambda alert logic to SNS notifications and Athena querying.
Built a distributed AWS backend integrating 6 services, triggered alerts at a 0.45 risk threshold, and structured Athena access over S3 datasets covering 140+ patient-day records.