Models are built using data. Most successful search, social media, and recommendation systems are built using personal models to provide people the right information, at the right time, in the right context, usually even before a user articulates his need. Food recommendation systems need to be built using the same approach. A personal food model is essential for recommending the right food item at the right time. It is also essential to predict the effect of food so the right suggestions can be made to avoid unpleasant situations. We need to build such models using food logs collected for the person. Many applications for foodlogging are being developed based on detecting the dish or item being consumed and finding nutritional elements based on ingredients in these items. Detecting items and the volumes consumed requires a multimodal platform and nutritional data sources for items prepared using specific ingredients and recipes. To see more information please watch this YouTube demo.