Presented By O'Reilly and Cloudera
Make Data Work
Dec 4–5, 2017: Training
Dec 5–7, 2017: Tutorials & Conference
Singapore

Train, Predict, Serve: How to go into production your Machine Learning model

Aki Ariga (Cloudera)
1:45pm2:25pm Wednesday, December 6, 2017
Data science and advanced analytics, Machine Learning
Location: Summit 1 Level: Intermediate

Who is this presentation for?

Data Engineers and/or Data Scientists

Prerequisite knowledge

Basic machine learning knowledge: how to build a predictive model, what ML pipeline is.

What you'll learn

Know the best practice of machine learning deployment patterns for production.

Description

Adopting machine learning system becomes an essential topic for enterprise companies to progress to the next stage of their business. However, there are many challenges for building machine learning system in production. Which architecture is best for this use case? How to deploy predictive models? How to avoid struggling with the difference between development and production environment?

Machine learning system tends to be a complex system, because it depends on different language, library, or framework, such as scikit-learn, TensorFlow, XGBoost. Knowing best practice will help you to choose right design pattern for going production.

In this talk, I will discuss what typical patterns of ML system are and how we can choose the best one. To apply best practices to your analytics platform, I will present examples of Hadoop/Spark ecosystem with Cloudera Data Science Workbench, which helps Data Engineers to bring Data Scientist’s ML models into production.

Photo of Aki Ariga

Aki Ariga

Cloudera

Field data scientist who is interested in service development with machine learning and natural language processing. After researching spoken dialogue system and building the large corpus analysis system, developing predictive services such as recipe recommendation.
Loving OSS development and hosting several tech communities (Ruby, machine learning, Julia).

Leave a Comment or Question

Help us make this conference the best it can be for you. Have questions you'd like this speaker to address? Suggestions for issues that deserve extra attention? Feedback that you'd like to share with the speaker and other attendees?

Join the conversation here (requires login)