Recommendations to others considering KITRUM:
KITRUM has a strong AI-skilled team with deep knowledge of the USA, European, and even Scandinavian markets. We recommend it for building mobile apps, web apps, design, and AI-powered solutions with full testing coverage and even project management support. Review collected by and hosted on G2.com.
What problems is KITRUM solving and how is that benefiting you?
Using Flutter with the Dart framework, we developed a cross-platform mobile application for iOS/Android. The application provides users with a fashion assistant by combining various technologies for wardrobe management and outfit generation. It integrated native device features, such as the Camera and Image Picker, for easy wardrobe input, and utilized Google ML Kit to perform on-device object detection for immediate clothing classification. For robust offline functionality, there are local storage solutions, such as Hive and SQLite, and managing the UI state reactively with Bloc or Riverpod. The backend architecture is built on Python using FastAPI to create a high-performance, containerized API. This setup allows for scalable processing of complex AI tasks, such as garment detection using YOLOX or Mask R-CNN models, and attribute classification with CLIP or EfficientNet. A PostgreSQL database manages the core application data, while Redis handles caching and background task processing with Celery. For advanced features like "Search Similar," a vector database solution, such as pgvector, Pinecone, or FAISS, can be used to enable fast, visual-based searches. The entire system is hosted on AWS, leveraging services like EC2 (GPU) for intensive model inference, S3 for media storage, and GitHub Actions for automated CI/CD pipelines. KITRUM also contributed to building a mobile app, design, project management, and QA. This comprehensive stack ensures a reliable, easily scalable, and personalized experience for users. Review collected by and hosted on G2.com.
