Productionising Machine Learning models - MLOps from a practitioners point of view

March 28, 2023 6:00 PM

Productionising Machine Learning models - MLOps from a practitioners point of view

On-demand
ENG
Registration
Technology stack
No items found.
Topic description

During the presentation, we will share learnings from multiple MLOps projects and show how we can efficiently build a reproducible, automated pipeline that follows the DevOps principles. We also will present common traps and risks with operationalizing ML projects while presenting an IoT use case.

The presentation answers the following questions:

  • How can I productionise my ML models?
  • How can I build an efficient development pipeline for ML models?    
  • How to build reproducible ML models?
  • How to track code, data, and parameter lineage with ML models?
  • What kind of deployment methods should I choose for my ML models?
  • What are the risks and the caveats of implementing MLOps?
  • How to apply ML models on edge? - an IoT use-case
Technology stack
No items found.
Summary

Building Machine Learning (ML) models, or AI models in popular culture, is increasingly common in all industries. As organizations progress from PoCs and the first projects with ML models that bring business value, they are faced with a big question: How can I put my model in production, how do I update it, and how do I close the gap between model development and production? In other words, how do I do MLOps?

event Partner(S)
Mndwrk
author
No items found.
Presenter(s)
Gulyás Máté
Mate Gulyas
CEO @DATAPAO
DATAPAO

Started his career as a backend developer but early on specialized in Data Engineering. Mate has more than 15 years of experience in Data Engineering and supporting Data Scientists to build enterprise ML solutions. Mate is currently the CEO of DATAPAO, a data consultancy company specializing in distributed systems, working closely with Databricks and Microsoft. He worked as a Principle Instructor and Practice Lead at Databricks. Before that, he was Co-Founder and CTO at enbrite.ly, an award-winning startup specializing in digital fraud detection.

Presenter(s)
Gulyás Máté
Gulyás Máté
CEO @DATAPAO
DATAPAO

Started his career as a backend developer but early on specialized in Data Engineering. Mate has more than 15 years of experience in Data Engineering and supporting Data Scientists to build enterprise ML solutions. Mate is currently the CEO of DATAPAO, a data consultancy company specializing in distributed systems, working closely with Databricks and Microsoft. He worked as a Principle Instructor and Practice Lead at Databricks. Before that, he was Co-Founder and CTO at enbrite.ly, an award-winning startup specializing in digital fraud detection.

REQUEST ACCESS
venue
Portrait of presenter

get in touch.

Tell us about your goals!
Contact Sales
Productionising Machine Learning models - MLOps from a practitioners point of view

March 28, 2023 6:00 PM

Mar 28, 2023

Productionising Machine Learning models - MLOps from a practitioners point of view

On-demand
ENG
Summary

Building Machine Learning (ML) models, or AI models in popular culture, is increasingly common in all industries. As organizations progress from PoCs and the first projects with ML models that bring business value, they are faced with a big question: How can I put my model in production, how do I update it, and how do I close the gap between model development and production? In other words, how do I do MLOps?

Partner(S)
Mndwrk
tags

A MndwrkOut órákon való részvételhez egyszeri regisztráció szükséges. Ha már közösségünk tagja vagy, nincs más dolgod, mint csütörtök 18:00-kor csatlakozni az e-mailben megkapott linken.

tags
March 28, 2023
Machine learning
Scikit learn
MLFlow
Databricks
Apache Spark
Docker
containerization
Azure
Azure IoT Hub
Presenter(s)
Gulyás MátéPortrait of presenter
Gulyás Máté
CEO @DATAPAO

Started his career as a backend developer but early on specialized in Data Engineering. Mate has more than 15 years of experience in Data Engineering and supporting Data Scientists to build enterprise ML solutions. Mate is currently the CEO of DATAPAO, a data consultancy company specializing in distributed systems, working closely with Databricks and Microsoft. He worked as a Principle Instructor and Practice Lead at Databricks. Before that, he was Co-Founder and CTO at enbrite.ly, an award-winning startup specializing in digital fraud detection.

Presenter(s)
Gulyás MátéPortrait of presenter
Mate Gulyas
CEO @DATAPAO

Started his career as a backend developer but early on specialized in Data Engineering. Mate has more than 15 years of experience in Data Engineering and supporting Data Scientists to build enterprise ML solutions. Mate is currently the CEO of DATAPAO, a data consultancy company specializing in distributed systems, working closely with Databricks and Microsoft. He worked as a Principle Instructor and Practice Lead at Databricks. Before that, he was Co-Founder and CTO at enbrite.ly, an award-winning startup specializing in digital fraud detection.

venue
Portrait of presenter
Topic description

During the presentation, we will share learnings from multiple MLOps projects and show how we can efficiently build a reproducible, automated pipeline that follows the DevOps principles. We also will present common traps and risks with operationalizing ML projects while presenting an IoT use case.

The presentation answers the following questions:

  • How can I productionise my ML models?
  • How can I build an efficient development pipeline for ML models?    
  • How to build reproducible ML models?
  • How to track code, data, and parameter lineage with ML models?
  • What kind of deployment methods should I choose for my ML models?
  • What are the risks and the caveats of implementing MLOps?
  • How to apply ML models on edge? - an IoT use-case
Registration

Regisztrációs listánk megtelt!

Registration

Regisztrációs listánk megtelt!

Registration

You can watch the video after registration

Az előadást regisztráció után tudod visszanézni

Open Positions
No items found.

Inspiring projects are waiting for you.

projects

March 28, 2023 6:00 PM

Productionising Machine Learning models - MLOps from a practitioners point of view

On-demand
ENG
Summary
összefoglaló
Registration
Regisztráció
venue
Portrait of presenter
Partner(S)
Partner
Mndwrk
WHAT IS HAPPY HOUR?
Mi az a happy hour?

Our HappyHour events offer a dedicated introduction to our key partners. Our goal with this event series is to give you a deep insight into a company before a possible formal interview.

This time, the spotlight will be on McKinsey's managers, who will be there to answer any questions you may have about the company. During the first 10-15 minutes of the event, a development manager will introduce the current team, the technologies, and the project, after which interested parties can ask any questions they may have.

agenda

During the presentation, we will share learnings from multiple MLOps projects and show how we can efficiently build a reproducible, automated pipeline that follows the DevOps principles. We also will present common traps and risks with operationalizing ML projects while presenting an IoT use case.

The presentation answers the following questions:

  • How can I productionise my ML models?
  • How can I build an efficient development pipeline for ML models?    
  • How to build reproducible ML models?
  • How to track code, data, and parameter lineage with ML models?
  • What kind of deployment methods should I choose for my ML models?
  • What are the risks and the caveats of implementing MLOps?
  • How to apply ML models on edge? - an IoT use-case
Registration
Regisztráció
Presenter(s)
Előadók
Gulyás MátéPortrait of presenter
Gulyás Máté
CEO @DATAPAO

Started his career as a backend developer but early on specialized in Data Engineering. Mate has more than 15 years of experience in Data Engineering and supporting Data Scientists to build enterprise ML solutions. Mate is currently the CEO of DATAPAO, a data consultancy company specializing in distributed systems, working closely with Databricks and Microsoft. He worked as a Principle Instructor and Practice Lead at Databricks. Before that, he was Co-Founder and CTO at enbrite.ly, an award-winning startup specializing in digital fraud detection.

Inspiring projects are waiting for you.

projects

Inspiring projects are waiting for you.

projects

Become a Mndwrk-er

Work, learn, and have fun together
Create your account
No items found.