Context and goal
During an AI software engineering internship at Training Edge Consulting, I developed a web application that turns conversational text prompts into generated images. The application also keeps conversations and generated images available to their owners.
Architecture
Django handles authentication, conversations, messages, and image records. Server-rendered HTML, CSS, and JavaScript provide the interface. A service layer calls the Stability AI API for image generation, while a translation service can convert prompts before generation. SQLite is the local database described in the repository.
Engineering decisions
- Conversations and generated images are associated with authenticated users.
- Generation settings, including image size, guidance scale, and inference steps, are configurable.
- Prompt validation and explicit handling of authentication, rate-limit, and upstream service errors keep failure states understandable.
- API credentials are read from environment configuration rather than embedded in the interface.
Outcome and limits
The repository contains the application and setup instructions. Image generation requires a Stability AI API key and a running service; this portfolio does not embed a live demo.