Productionizing Machine Learning with Microservices Architecture

Explore a 35-minute talk on productionizing machine learning models using microservices architecture. Learn how to streamline the process of moving workloads from training to production by running Spark as a microservice for inferencing. Discover techniques for achieving auto-scaling, versioning, and security in machine learning deployments. Gain insights into feeding feature vectors aggregated from multivariate real-time and historical data to machine learning models and serverless functions for real-time dashboards and actions. Delve into topics such as machine learning pipelines, service composition, workflow automation, microservices architecture, monitoring, and running Spark in distributed clusters. Understand how to create larger pipelines, define domain-specific languages, and implement automated workflows for efficient machine learning operations.

Intro
Machine Learning Pipeline
Service
Composition
Workflow
Automated workflows
Payoneer
Traditional Active Architecture
Microservices Architecture
Monitoring
ML Run
ML Project
Code to Function
Analyzing Data
Running a Job
Running on a Distributed Cluster
Creating a Bigger Pipeline
Defining a DSL
Automated workflow
Spark