MLOps, Machine Learning Operations for beginners

Name: MLOps, Machine Learning Operations for beginners
Rating: 3.9 (15 reviews)

Machine Learning models from experimentation to production

Created byDat Art

Last updated 8/2022

English

English [Auto],

What you'll learn

Understand the lifecycle of a Machine Learning model
Gain the best practices for putting Machine Learning models in production
Leverage the power of MLOps to productionalise Machine Learning models at scale
Get some insights on how to choose your perfect MLOps stack

Course content

4 sections • 16 lectures • 51m total length

Course Introduction2:30
Discover the fundamentals of MLOps, compare it to DevOps, and explore the machine learning lifecycle, key actors, and how to shape your ML stack for effective operations.
Course audience and prerequisites1:19
Identify the audience from data scientists to software engineers, IT professionals, and DevOps, who want to productionize and deploy machine learning models. Accept optional prerequisites; some experience helps.
Take the most of this course1:08
Learn how to take the most of sports and technical courses through an interactive format, using the Q&A section, playback speed controls, and star ratings.

MLOps: ML and Ops3:04
Define mlops as the blend of machine learning and operations, where a model is an artifact from the learning process used for predictions, and ops sustains reliable infrastructure.
MLOps in the eyes of the giants2:31
Compare how giants define envelopes in MLOps across GCP, AWS, and Microsoft, highlighting DevOps, automation, monitoring, and governance for machine learning models in release management.
From MLOps to DevOps and Vice-versa4:03
Explore DevOps as a paradigm uniting development and operations to accelerate deployment, with QA and security teams, guided by collaboration, automation, continuous improvement, and consumer centric actions for MLOps.
Traditional Vs. Machine Learning programming - Part 17:30
Compare the lifecycle of classical software with machine learning models, highlighting training data in development. Track metrics, parameters, and model drift through development, testing, deployment, and monitoring for continuous feedback.
The Machine Learning Lifecycle - Part 17:14
Explore the machine learning lifecycle from framing a business question to data exploration, feature engineering, experimentation, and evaluation, emphasizing iterative development and collaboration between data scientists and data engineers.
The Machine Learning Lifeycyle - Part 23:29
Trace the machine learning lifecycle from development to deployment, covering registry, feature store readiness, go/no-go feasibility, deployment modes, and drift-aware performance monitoring.

Subject Matter Expert1:59
The subject matter expert identifies problems, translates business needs to data scientists, and sets KPIs, while ensuring lifecycle collaboration and clear feedback among data scientists and non-technical stakeholders.
Data Scientist3:03
The data scientist aligns business needs with data collection and collaborates with experts to guide model development through exploration, feature engineering, and evaluation, delivering an operational production-ready model.
Data Engineer1:51
Data engineers build data pipelines to ensure quality and availability for training and model development, collaborating with data scientists and subject matter experts to apply transformations and trace lineage.
Software Engineer1:14
Software engineers integrate models into existing applications, ensuring compatibility with IT frameworks and infrastructure, and collaborating with scientists to adapt or refactor the model.
DevOps Engineer1:14
Define the DevOps engineer role in integrating machine learning model lifecycles into existing CI/CD pipelines, building secure operational systems, and delivering seamless integration with standard DevOps practices.
Machine Learning Engineer2:18
Discover the role of a machine learning engineer, bridging data science and software engineering by refactoring notebooks, monitoring models, provisioning infrastructure, and handling model drift within end-to-end ml systems.

Requirements

Basic understanding of Machine Learning concepts (preferred)
Previous Experience in Software Development (preferred)
Some Experience in DevOps (not mandatory)

Description

This course is about Machine Learning Operations.

Machine Learning and Artificial Intelligence have became a hot topic in recent years. Numerous techniques and algorithms were developed and proved their efficiencies in addressing business issues and bringing value to companies. Take fraud detection, recommendation systems or autonomous vehicles, etc. as examples.

However, most of the developed machine learning models do not go to production! Among others, this is due to one major reason: Machine Learning models are not classical software. The existing frameworks and methodologies that work for classical software proved to be inadequate with Machine Learning models. Hence, new paradigms and concepts should be brought to handle the specificities of Machine Learning Algorithms.

This course is addressed to Data professionals (Data Scientists, Data Engineers, Machine Learning Engineers and Software Engineers) as well as to everyone who want to understand the lifecycle of a Machine Learning model from experimentation to production. In this course, wa re going to see the best practices and recommended ways to put machine learning models into production. This will allow us also to see how we can leverage the power of MLOps to deploy Machine Learning at scale. Finally, as deploying models is about tooling, we are going to have a look on how to choose its perfect stack when adopting Machine Learning Operations best practices.

Wish you a nice journey!

Who this course is for:

Everyone
Data Scientists
Machine Learning Engineers
Software Engineers
DevOps Engineers

MLOps, Machine Learning Operations for beginners

What you'll learn

Explore related topics

Course content

Introduction3 lectures • 5min

MLOps Concepts6 lectures • 28min

MLOps actors6 lectures • 12min

MLOps tools1 lecture • 7min

Requirements

Description

Who this course is for: