Use state-of-the-art functional programming that enables large scale data collection and processing. Let it exploit the fine and coarse grain parallelism that modern day processor architectures have in abundance. Use meta-programming, powerful macros and Domain Specific Languages (DSLs). Re-use design patterns such as GenServer and benefit from scalable out-of-the-box ETS caching.

Automatically and resiliently schedule difficult to predict network latencies. Efficiently stream and process data through lazy evaluation. Become fault tolerant by supervising execution. Be agile and reduce maintenance overhead. Distribute data at low latency and let applications run and fly at scale.

Get ready for Machine Learning. Develop the data models, collect the necessary data, train/adjust/retrain the models and enrich the data when and where necessary to positively impact and extend product, service or operation. Improve the business bottomline.

Machine Learning makes the AI algorithms it generates model generic and data training specific. Before applying a trained AI algorithm we must ensure that its training data reflects the real world in which the AI will be applied. This puts continuous capture of proper representative training data as well as appropriate test data at the heart of maintaining algorithmic integrity of Machine Learning derived intelligence as general applicable AI.

Develop and share knowledge, offer participation and build understanding among and beyond your own team.

Enrich your daily workflow with code and data narratives that combine prose, code source, inline code execution and data visualization. Narratives that can be documented so they become self explanatory. Naratives that telecommute particular issues of concern with anybody nearby or remote for call to action, advice, on-boarding or simply sharing. With everybody being literally on the same page!

An expressive and flexible code & data workflow shared within the Machine Learning pipeline makes the process of Machine Learning and its derived AI intelligence much more transparent. This is key to a deeper understanding of data, a continuous improvement of Machine Learning and getting the best possible results as AI.

Web enable your AI application with Web server, Pub/Sub message server, low latency Web sockets, liveview Web page roundtrips and a database backend. Get ready for scale by using efficient data caching besides geo-located application proxies with a local database replica that connects to its database master through a dedicated high speed IPv6 backbone.

Build rich, interactive web applications quickly, with less code and fewer moving parts. Join the growing community of developers using Phoenix to craft APIs, HTML5 apps and more, for fun or at scale.

Set out and don't be afraid to miss. Practise to close in. And increasingly learn to hit your AI goals!

Target the Machine Learning generated AI intelligence to its business use by its business participants. Let the business monitor effectiveness by measuring true/false positives and true/false negatives. Feed these ratios back to the Model & Train stage of the Machine Learning pipeline.

Involve the business operational department that can make sound business judgements within their own specialized expertise be it sourcing, production, legal, finance, admin, human resources, publicity or marketing & sales.