we will have a new meetup on June 5th and I am happy to have the agenda ready and the guys and ladies from Lentiq already signed up to come and tell us about how to productionize machine learning. Thanks Cosmin and Costina for taking the time to present in our meetup. Also big thanks to Felix for bringing his perspective on blockchain and big data in our meetup.
6:30 - 7:00 PM Arrival
7:00 - 7:40 PM Productionizing Machine Learning, Cosmin Pintoiu & Costina Batica @ Lentiq (https://www.linkedin.com/in/cosmin-pintoiu/ ,https://www.linkedin.com/in/costina-b%C3%A2tica-2578b911b/)
Usually a machine learning project ends at evaluating the model: Do we have acceptable accuracy? Good. Now what?! The value of a ML model though is making accurate predictions for the users, deploying and maintaining the model should be easy and automated. But how?
We will answer these questions by:
- preparing Jupyter notebooks for production as reusable code blocks
- creating an ML workflow
- and finally serving online, at scale predictions using Lentiq Model Server.
Along the way we will use: Lentiq, Spark, Jupyter, MLeap, Flask, Kubernetes.
7:40 - 8:15 PM Technology showdown: database vs blockchain, Felix Crisan, Netopia ( https://www.linkedin.com/in/fixone/ )
Blockchains and Big Data play well together. One can look at a blockchain like just a new kind of distributed database and, at the same, realize that the trust minimising assumptions present in these systems can unlock additional value for ‘regular’ big data projects – either by offering an alternative to centralized storage or by easing the process of providing data and training models.
Look forward seeing you in June,