HR, ERP, Clinical, Analytics, and Monitor modules run independently on Azure App Service.
Enterprise data platform portfolio
Meridian operating system for analytics workflows.
A cloud-hosted portfolio demonstrating operational apps, event-driven data generation, Azure Data Lake ingestion, Snowflake/dbt modeling, embedded Power BI analytics, LangGraph-assisted analysis, and end-to-end data engineering monitoring.
Architecture overview
Operational apps produce events, analytics systems turn them into decisions.
HR, ERP, and Clinical are independent Azure-hosted applications. Each writes event envelopes into Azure Data Lake and exposes current operational APIs. Snowflake stores the historical analytics warehouse, dbt creates integrated models and marts, Power BI visualizes the curated semantic layer, and LangGraph answers governed natural-language questions over the same marts.
Azure Data Lake stores raw JSON events. Azure SQL stores current operational app data.
Snowflake, dbt, Power BI, and LangGraph support reporting, semantic models, and AI-assisted analysis.
Data flow diagram
Portfolio data flow
1 HR / ERP / Clinical
Users create workflow and domain events.
2 Azure App Services
APIs serve apps and write event envelopes.
3 Azure Data Lake
Raw JSON events land in ADLS paths by app/date.
4 Snowflake + dbt
RAW, STAGING, INTEGRATION, MARTS, and PBI views.
5 Power BI + LangGraph
Executive dashboards and natural-language analysis.
6 Monitor
Checks apps, jobs, dbt, Snowflake, reports, and chatbot health.
How to evaluate the site
Suggested review path
Open HR, ERP, and Clinical. Submit a workflow form or inspect the seeded current data panels.
Open Analytics, load the embedded Power BI report, and ask the assistant: show total events by app.
Open Monitor and review app health, job telemetry, Snowflake/dbt status, Power BI checks, and chatbot checks.
Technology stack summary
What is demonstrated
App Service, Container App Jobs, Azure SQL, Data Lake Storage, managed identities, RBAC, and Key Vault integration.
Python, JSON event envelopes, Snowflake, dbt, data quality tests, persisted dbt artifacts, and warehouse observability.
Power BI Embedded, semantic views, LangGraph chatbot, role-scoped analytics queries, and executive reporting.
Known limitations
Current portfolio constraints
Analytics role hardening was deferred so reviewers can test the public site without sign-in friction.
The HR, ERP, and Clinical events simulate an Evergen-style global workflow scenario for portfolio-scale reporting.
Monitor already tracks apps, Azure jobs, Snowflake/dbt, Power BI, and chatbot probes. Future work can add alerting and richer Power BI capacity telemetry.
Resources are sized for portfolio demonstration and may be scaled differently in a production environment.