VitalConnect securely automates thousands of cardiac reports daily with Loka on AWS
Industries
Healthcare,
Remote Patient Monitoring
Tech & TOOLS
AWS Lambda, Amazon SQS, Amazon ECS Fargate, Amazon SageMaker Real-Time Inference, SageMaker MLflow, Amazon S3, Amazon CloudWatch, AWS WAF
Teams & Services
AI & ML, Cloud Engineering, DevOps, Data Engineering, Security & Compliance
milestones
4 weeks to real-time patient reporting, FDA-regulated workflows, production-grade security, HIPAA-compliant AWS infrastructure
Industry
Healthcare,
Remote Patient Monitoring
Tech & TOOLS
AWS Lambda, Amazon SQS, Amazon ECS Fargate, Amazon SageMaker Real-Time Inference, SageMaker MLflow, Amazon S3, Amazon CloudWatch, AWS WAF
Teams & Services
AI & ML, Cloud Engineering, DevOps, Data Engineering, Security & Compliance
milestones
4 weeks to real-time patient reporting, FDA-regulated workflows, production-grade security, HIPAA-compliant AWS infrastructure
VitalConnect delivers automated daily reports in-house for thousands of patients with secure, real-time cardiac monitoring infrastructure.
The situation
VitalConnect is a healthcare technology company specializing in wearable biosensors and remote cardiac monitoring. Its platform processes real-time ECG data to help clinicians monitor patients continuously and identify cardiac events that require attention.
VitalConnect needed to bring its analytics and reporting capabilities in-house to strengthen control over patient data, support FDA compliance, and reduce reliance on third-party analytics and reporting vendors.
The company partnered with Loka to design and implement a secure AWS platform capable of ingesting continuous ECG streams, running real-time machine learning inference, supporting technician adjudication, and automatically generating clinician-ready daily and end-of-study reports.
The challenge
VitalConnect needed to replace external analytics and reporting services while maintaining reliable processing of continuous cardiac data from tens of thousands of patients.
The new platform needed to process real-time ECG streams, detect anomalies, support technician review, and generate reports without creating manual bottlenecks. It also needed production-grade security and compliance controls appropriate for protected health information and FDA-regulated workflows.
VitalConnect also needed a repeatable path from ML development to production. Model training, deployment, monitoring, release procedures, and model lineage needed to be standardized across the ML, data, and application layers.
The platform had to support controlled production releases, including blue/green deployments and audited manual release procedures for the operations team.
The solution
Loka led a Design Sprint and Discovery process to define daily and end-of-study reporting requirements and establish the path from development through production.
Loka designed an AWS architecture for streaming ingestion, real-time ML inference, technician adjudication, and automated report generation. AWS Lambda, Amazon SQS, and Amazon ECS Fargate support the processing pipeline, while Amazon SageMaker provides managed real-time ML inference.
Loka also refactored ML training into a managed SageMaker and MLflow workflow with CI/CD and model lineage, creating a more consistent path for developing, tracking, and deploying models.
To strengthen operational control, Loka established blue/green deployment strategies across API, web, ML, and ingestion services. A controlled runner environment enables VitalConnect's operations team to execute manual, audited releases and standardized rollbacks.
Security and compliance controls were built into the architecture, including Amazon Cognito, AWS WAF, AWS Secrets Manager, Amazon CloudWatch, and AWS-native network and access controls. The resulting infrastructure provides a HIPAA-compliant foundation designed to maintain 99% or higher uptime.
What we delivered
Automated cardiac monitoring reports
4 weeks: Loka delivered automated daily reporting, including zero-touch generation and delivery to VistaCenter, replacing manual and third-party reporting processes.
Real-time ML inference
Continuous ECG streams: Loka designed infrastructure to ingest and process real-time cardiac data, run ML inference, detect anomalies, and support clinician-ready reporting.
In-house reporting platform
1 platform: Loka replaced external analytics and reporting vendors with an in-house pipeline, giving VitalConnect greater control over data, compliance, and reporting workflows.
Production ML operations
End-to-end ML workflow: Loka implemented managed SageMaker and MLflow workflows with CI/CD and model lineage, creating a controlled path from model training through deployment.
Secure, controlled releases
4 deployment surfaces: Blue/green release strategies were established across API, web, ML, and ingestion services, with standardized rollback procedures to reduce deployment risk.
Technician adjudication
Tracked review workflows: Loka added adjudication tracking and analytics to support technician productivity and provide additional visibility into model performance.
Project results
VitalConnect moved from reliance on external analytics and reporting services to an in-house AWS platform capable of processing real-time cardiac data and automatically delivering daily reports. By combining real-time ML inference, automated reporting, managed ML workflows, controlled deployments, and HIPAA-compliant AWS infrastructure, Loka helped VitalConnect create a production-ready platform designed to support continuous monitoring for tens of thousands of patients.
In 4 weeks: Delivered automated daily reports, including zero-touch generation and delivery to VistaCenter.
Thousands of patients: Built infrastructure designed to process continuous cardiac monitoring data across a large remote patient population.
99%+ uptime: Designed a production infrastructure foundation to support high availability for critical cardiac monitoring workflows.
Real-time processing: Enabled ECG streams to move through ingestion, ML inference, anomaly detection, technician adjudication, and reporting workflows.
In-house control: Replaced third-party analytics and reporting services, giving VitalConnect greater control over patient data, compliance, and reporting.
Repeatable releases: Standardized blue/green deployments and audited release procedures across API, web, ML, and ingestion services.
Production ML: Connected model training, lineage, CI/CD, and real-time inference through managed SageMaker and MLflow workflows.
