BrightInsight Regulated Digital Health Platform
Problem Space
Build and maintain a global regulated SaaS platform for digital health products with multi-cloud deployment (AWS, Azure, GCP) and regulatory compliance. Enable secure-by-default data handling across biopharma and medtech partners. Integrate early-stage ML/LLM capabilities for predictive clinical models while maintaining composability across 8 global engineering teams.
Architecture & Patterns
- Modular microservice architecture spanning AWS, Azure, and GCP — secure-by-default
- Self-describing APIs and composable platform modules
- Integrated early-stage LLMs for predictive models for patient deterioration
- Embedded contract testing (Pact) across 8 global teams — composable contracts let specialist teams move independently without a coordination tax
- CI/CD pipelines with DevSecOps gates via GitHub Actions, Terraform, and Ansible
- ML model training and tracking with PyTorch, TensorFlow, MLflow, and Delta Lake
Tools & Stack
Go, Node.js, TypeScript, C#, Pact, Terraform, Ansible, GitHub Actions, AWS, Azure, GCP, PyTorch, TensorFlow, MLflow, Delta Lake
Business Outcomes
- 25% improved detection of patient deterioration via early-stage LLM predictive models
- Over 20 million API calls per day across the platform
- Embedded contract testing across 8 global teams reduced integration failures
- Strengthened audit readiness via ISO 13485 and IEC 62304 mapping
Regulated / Domain Context
ISO 13485, IEC 62304 — medical device software lifecycle and quality management
Reusable Narrative Snippets
Built a global health data platform spanning AWS, Azure, and GCP with modular microservice architecture and secure-by-default design, serving over 20 million API calls per day.
Integrated early-stage LLMs for predictive models that improved patient deterioration detection by 25%.
Embedded contract testing with Pact across 8 global teams and introduced self-describing APIs and composable platform modules.
Source Notes
- Derived from role responsibilities and achievements in
config/madu_profile.json; reconciled with JobVia export (madu_alikor_export.json). - Confidence: high