Bringing DevOps to Data Science: Operationalize AI Leveraging Postgres
No video of the event yet, sorry!
(Greenplum Technical Session for PostgresConf Wednesday
Successful enterprise AI applications in 2018 are ecosystems of machine learning solutions that tightly integrate a feedback loop triggering automated updates to the underlying algorithms - creating closed loop machine learning systems.
In order to efficiently build and scale these systems enterprises need reliable, highly performant and extensible data tools to not only wrangle and prepare complex disjoint structured and unstructured data but also build and deploy machine learning algorithms. With the diversity of the Postgres community projects including PostGIS for Geospatial, Apache MADlib Postgres based machine learning, Massively Parallel Postgres analytics engine Greenplum for big data, procedural language extensions to the Python and R package ecosystems and now MADlib Flow for containerized deployment of AI pipelines to Kubernetes, Postgres provides one of the most compelling software solution stacks for enterprise AI available today.
During this 50-minute breakout session, the presenters will:
Highlight the advantages of Postgres and Postgres community projects for enterprise AI
Recommend a deployment template for closed loop machine learning solutions using Postgres community projects
Provide a pre-release preview of the MADlib Flow ML pipeline deployment project
Showcase the art-of-possible with Postgres as an enterprise AI software solution with a demo of a real time financial transaction fraud prevention system - built using Greenplum, MADlib and Kubernetes - that continuously learns new threat signatures and can scale to handle a high transaction throughput and low latency response requirements
- 2019 March 20 14:00
- 50 min
- Sutton Place
- Postgres Conference