Marius-Alexandru A Voinescu

I designed and implemented a scalable IAM platform using Terraform to standardize application onboarding and access control across a multi-tenant hub-and-spoke architecture in Okta. The solution automates the creation of SAML and OIDC integrations in a central hub tenant while dynamically provisioning access pathways (via bookmark applications) across 10+ spoke tenants.

I developed reusable, parameterized Terraform modules that abstract complexity and allow application definitions to be driven via structured inputs (e.g., JSON), supporting multiple authentication types, tenant scoping, and environment-specific configurations. This significantly reduced manual configuration effort and ensured consistent enforcement of security controls.

The platform integrates with Terraform Cloud to introduce approval gates, speculative plans, and drift detection, providing full auditability and controlled change management. This approach improved deployment consistency, reduced configuration drift, and strengthened access governance across distributed environments.

This project highlights my expertise in IAM architecture, infrastructure as code, and building secure, scalable identity platforms aligned with enterprise security standards.

I designed and deployed a self-hosted security and observability platform across multiple nodes using Terraform-managed infrastructure. The environment includes Wazuh (SIEM), Grafana, and centralized logging to monitor system activity, detect anomalies, and provide security visibility across services.

The architecture leverages Docker for service orchestration and Tailscale for secure connectivity, ensuring no direct exposure of internal services. Infrastructure is fully managed as code, enabling reproducibility, version control, and controlled deployments.

In addition, I am implementing automated incident response workflows using N8n by integrating Wazuh alerting with a self-hosted project management system (Plane). High-severity alerts and specific security events trigger automated ticket creation, enabling structured incident tracking and response handling. This bridges detection with action and simulates real-world SOC workflows.

This project demonstrates hands-on experience with SIEM deployment, alerting pipelines, infrastructure as code, and building practical security operations processes in a controlled environment.

Chronicle is a Go-based data ingestion and normalization platform designed to transform unstructured ISBN inputs into enriched, analytics-ready datasets. The system serves as the backbone for a book intelligence platform, integrating multiple external sources (Google Books, Open Library, and planned scraping pipelines) to standardize and enrich metadata.

I designed Chronicle with a modular, extensible architecture focused on performance and clean separation of concerns. Core capabilities include ISBN normalization (handling inconsistent formats and ISBN-10/13 conversion), database ingestion (PostgreSQL), and a planned enrichment layer leveraging LLMs to extract themes and contextual insights.

The platform integrates with a Grafana dashboard (provisioned via Terraform), enabling visualization of reading behavior, genre distribution, and trend analysis. This reduces manual data handling effort and enables near real-time analytics on structured reading data.

This project demonstrates end-to-end system design across back-end engineering, data pipelines, and observability, with a strong emphasis on scalability and maintainability.