Turning Complexity
Into Systems

Nomia is a Python CLI tool for tracking the relationship between business rules and their implementation in code.

Instead of letting domain knowledge drift across old tickets, scattered conditionals, and outdated documentation, Nomia makes rule-to-code alignment explicit and observable. Rules are declared in a YAML file, linked to implementation functions with decorators, and validated through CLI commands.

The workflow centers on three commands: validate creates a snapshot of current rule implementations, check compares code against the last validated state, and audit helps discover functions that are not linked to any rule.

Nomia is designed to fit into real development workflows such as pull requests and CI pipelines. It does not replace tests or infer business logic automatically; it gives teams visibility into whether code still reflects the business intent they chose to track.

Projects API is a backend engineering study focused on clean architecture, domain rules, and layered application design.

The API manages software project data and exposes public endpoints for retrieving published portfolio projects, along with authenticated admin endpoints for maintaining project records. It supports language filtering, pagination, publication workflows, and a health check endpoint.

The implementation separates FastAPI routes, application services, repository abstractions, and transaction handling through a Unit of Work pattern. This keeps domain logic testable independently from the web framework and database layer.

The project also includes a reproducible development environment with Docker and Dev Containers, plus CI checks for code quality, testing, and security.

Plant Spy is a backend simulation of an industrial monitoring system inspired by real-world asset monitoring platforms.

The system models the lifecycle of industrial monitoring: components generate sensor readings, readings are evaluated for anomalies, repeated anomalies become alerts, alerts escalate in severity, and high-severity alerts generate work orders.

Plant Spy is designed as a modular monolith using an in-memory EventBus and pub/sub communication. Modules publish and react to events instead of directly calling each other, keeping responsibilities separated while still running inside one backend application.

The module boundaries cover hierarchy, readings, processing, alerts, and work orders. The project emphasizes event-driven design without external brokers, clear backend boundaries, and the operational flow behind anomaly detection and maintenance escalation.