ML Model Deployment and Scoring on the Edge with Automatic ML & DF / Flink2Kafka

image
Details

6:00 PM (EDT)

Machine Learning Model Deployment and Scoring on the Edge with Automatic Machine Learning and Data Flow

Deploying Machine Learning models to the edge can present significant ML/IoT challenges centered around the need for low latency and accurate scoring on minimal resource environments. H2O.ai's Driverless AI AutoML and Cloudera Data Flow work nicely together to solve this challenge. Driverless AI automates the building of accurate Machine Learning models, which are deployed as light footprint and low latency Java or C++ artifacts, also known as a MOJO (Model Optimized). And Cloudera Data Flow leverage Apache NiFi that offers an innovative data flow framework to host MOJOs to make predictions on data moving on the edge.

Join us for an interesting talk and learn how to push Driverless AI Machine Learning models to the edge using Cloudera Data Flow.

We will be using a public UCI ML Hydraulic Systems Condition Monitoring dataset to build a Driverless AI model, download it as a MOJO scoring pipeline and push it to the edge using Cloudera Flow Management (Apache NiFi). This meetup will be driven around this scenario from slides to demo. We will also show how to do ML edge inference with Cloudera Edge Management (Apache MiNiFi C++). In this use case, ML edge inference can be used for Hydraulic System Predictive Maintenance for monitoring the Hydraulic cooling condition, valve condition, pump leakage, accumulator, and stability.

Speakers:

James Medel (H2O.ai - Technical Community Maker)
Greg Keys (H2O.ai - Solution Engineer)
Tim Spann (Cloudera)

FLaNK Stack Integration with MiNiFi, NiFi, Kafka and Flink.

7:00 PM (EDT)

Kafka 2 Flink - An Apache Love Story

This project has heavily inspired by two existing efforts from Data In Motion's FLaNK Stack and Data Artisan's blog on stateful streaming applications. The goal of this project is to provide insight into connecting an Apache Flink applications to Apache Kafka.

Ian R Brooks PhD | Senior Solutions Engineer & Data | Cloudera

Web Conference information to be provided.

James Medel (Moderator)

Engineer & Community Maker at H2O.ai

Greg Keys (Moderator)

Solution Engineer at H2O.ai

Tim Spann (Moderator)

Principal DataFlow Field Engineer at Cloudera

Ian Brooks (Moderator)

Senior Solutions Engineer & Data Scientist at Cloudera

Key:

Complete
Failed
Available
Locked
Meetup Section 1 (Replay): ML Model Deployment Overview
Click on view to access the replay, the slides and see the description of the video.
Meetup Section 2 (Replay): Mojo Deployment to Apache NiFi Demo
Click on view to access the replay, the slides and see the description of the video.
Meetup Section 3 (Replay): Questions on Mojo Deployment
Click on view to access the replay and see the description of the video.
Presentation Slides (PDF): ML Model Deployment and Scoring on the Edge with Automatic ML & DF
Open to download resource.
Open to download resource. Slides cover ML Model Deployment using Driverless AI MOJO and Cloudera DataFlow
Presentation Slides (PDF): Flink & Kafka
Open to download resource.
Open to download resource. Slides cover Flink & Kafka The original slides of the content owner Ian Brooks are located on his GitHub: https://github.com/BrooksIan/Flink2Kafka/blob/master/Future%20of%20Data%20-%20Flink%20%26%20Kafka.pdf
Meetup Section 4 (Replay): Transition to Apache Flink
Click on view to access the replay, the slides and see the description of the video.
Meetup Section 5 (Replay): Flink Reference Architecture App Demo
Click on view to access the replay, the slides and see the description of the video.
Meetup Section 6 (Replay): Questions on Real-Time Streaming ML
Click on view to access the replay, the slides and see the description of the video.
You must be logged in to post to the discussion
  • RC

    I see that tony got in.

    Reply
  • RC

    The button should be under the contents tab and say something like open presentation

    Reply
  • RC

    It seems like the direct zoom link is not allowing folks to get in.  Please login, register and then click the button to open live stream.

    Reply
  • RC

    Are folks login into the H2O.ai training site?  Once you login and register for the meetup you will see an option to open the live stream.

    Reply
  • HD

    Get the same error, looks like this:

    Zoom API Error: User does not exist: hd227@scarletmail.rutgers.edu.

    Sorry, you do not have permission to access /webinar/187469 with Request method: GET

    Reply
  • TH

    same error here.

    Reply
  • NM

    Is anyone reading this?

    Reply
  • NM

    Hi! I'm having some trouble to join. Link in "Contents" redirects to a page with error: "Zoom API Error: User does not exist"

    Reply
  • TH

    Hi, I just registered this, but I don't see any place I can get in, can you advise?

    Reply
  • RC

    We are now live!

    Reply