01. Roadmap for eShopOnContainersAI (Artificial Intelligence features)

Current features - v1.0 Alpha

Product recommendations based on Azure Machine Learning Studio models
Products search based on image recognition through multiple approaches:
- Azure Cognitive Services (Easy/basic Computer Vision approach)
- TensorFlow + TensorFlowSharp (Advanced Deep Learning approach)
  - Using a pre-trained model
  - Using a custom model to be trained with Azure ML Workbench (Vienna)
- CNTK (Advanced Deep Learning approach)
  - Currently, the CNTK microservice needs to be deployed on Windows / Windows Container since CNTK still doesn't support .NET Core.
BOT as client app for eShopOnContainers
- Surfacing AI features like product search based on image recognition
- Language understanding (based on L.U.I.S.) to access eShop features like Order's status

Internal Dashboard app with Orders Predictions based on a model trained with Azure ML Workbench (Vienna)
TBD

Recommendation systems
Product recommendation (ML Studio, C#)
Computer Vision
Image classification
- Cognitive Services - Computer Vision
- Congnitive Services - Custom Vision (C#)
- Cognitive Services - Mobile applications (Xamarin, C#)
- TensorFlow - Pre-Trained models (C#)
- TensorFlow - Custom Model (ML Workbench, CNTK, TensorFlow, C#)
Natural Language Processing
Skype Bot (Microsoft Bot Framework, LUIS, C#)
Regression
Sales Forecasting (ML.NET, C#)