The Hive Think Tank

Document created by maprcommunity Employee on Oct 4, 2017Last modified by maprcommunity Employee on Oct 12, 2017
Version 2Show Document
  • View in full screen mode


Date:  Monday, October 23, 2017
Location:  Silicon Valley Bank- 3005 Tasman Dr., Santa Clara 95054
Time: 6:00pm
Registration Link:



How do you get a machine learning system to deliver value from big data in a real world setting?


Turns out that 90% of the effort required for success in machine learning is not the algorithm or the model or the learning – it’s the logistics. That 90% comes from many things, including the need to stage and deploy multiple versions of each model, to carefully collect and curate updated training data and to monitor model performance. Lately we have added scale, speed and the need to handle multiple machine learning frameworks at the same time to make the problem more difficult.


There is a way to make this easier and more effective – the rendezvous architecture. This new design for model and data management is based on a streaming approach in a microservices style. It makes use of containerization and orchestration to solve many of the problems involved in continuous deployment of machine learning models. In presenting the rendezvous architecture, Ted Dunning, our speaker for the night, will cover techniques for effective model-to-model comparisons, for model deployment and management in production- including importance of new models on stand-by, and model monitoring.  Finally, Ted will talk about how a DataOps style of organization matches the flexibility offered by the rendezvous approach for machine learning.



Ted Dunning - Chief Applications Architect at MapR




The Exchange

[Book Discussion] - Model Management in the Real World