A robust, secure, and scalable system developed for a digital voting platform. This project was built to ensure transparency, complex state management, and high availability during simulated voting processes, integrating modern UI/UX principles and AI capabilities.
The system was designed following a clean architecture pattern to ensure separation of concerns, testability, and maintainability. The core solution bridges a robust backend with a reactive frontend ecosystem:
- API / Web: The presentation layer handling secured HTTP requests, exposing RESTful endpoints, and delivering a reactive user interface (Blazor).
- Business Logic: Contains the core functionality, service layers (including AI integrations), and complex state machines governing the voting rules.
- Domain: The heart of the software, defining core entities, interfaces, and domain exceptions without external dependencies.
- Persistence: Responsible for data access, implementing patterns to interact with the underlying relational database.
- Language & Framework: C#, .NET, Blazor
- Architecture: N-Tier, REST API
- UI/UX: CSS, Global Components, Markdown Parsing
- Integration: LLM APIs (Artificial Intelligence)
This software was developed collaboratively by an engineering team. Within this cross-functional environment, my specific responsibilities and architectural implementations included:
- AI-Driven Roadmap Generator: Engineered a dedicated backend service to asynchronously query LLM APIs, dynamically constructing system prompts using historical user data (scores and feedback).
- Full-Stack Implementation: Built secured endpoints to serve the AI responses and developed a reactive frontend with state management (loading/success/error states) and secure Markdown-to-HTML parsing.
- Dynamic Role-Based Access: Designed a complex authorization system segregating Public Voting from Authenticated Jury access, synchronizing UI toggles with backend DTOs.
- Data Isolation: Implemented strict backend filtering for the Jury Dashboard to guarantee data privacy, ensuring unauthorized requests receive a
403 Forbiddenresponse.
- Lifecycle State Machine: Developed the manual execution logic for voting events (Start, Stop, Resume, Finish), strictly enforcing valid state transitions, dynamic timestamp adjustments (
DateTime.UtcNow), and transaction blocking during paused states. - Validation & Cascading Logic: Handled complex edge cases in data persistence, including restricted deletion protocols (preventing the unlinking of active events) and comprehensive timeline validation.
- Cognitive Load Reduction: Redesigned complex interfaces following Nielsen's Heuristics, refactoring standard continuous-scroll views into progressive Wizard patterns with Blazor state preservation.
- Component Standardization: Unanimously improved UI consistency by developing global alert components for predictable error handling and unifying CSS design tokens across the application.
To maintain high code quality and coordinate effectively across the team, we adopted industry-standard agile practices:
- Version Control (Git): We utilized a strict Git workflow, ensuring the
mainbranch always contained production-ready code. - Pull Requests & Code Reviews: All new features and bug fixes were developed in isolated feature branches. Code was merged into the main branch exclusively via Pull Requests, requiring peer review to maintain architectural standards.