Machine Learning in Production
Keep adding new skills with 10,000+ programs for $239 (usually $399). Save now.
3,363 reviews
Recommended experience
3,363 reviews
Recommended experience
What you'll learn
Identify key components of the ML project lifecycle, pipeline & select the best deployment & monitoring patterns for different production scenarios.
Optimize model performance and metrics by prioritizing disproportionately important examples that represent key slices of a dataset.
Solve production challenges regarding structured, unstructured, small, and big data, how label consistency is essential, and how you can improve it.
Skills you'll gain
- Continuous Deployment
- Data Integrity
- MLOps (Machine Learning Operations)
- Model Optimization
- System Monitoring
- Machine Learning
- Continuous Monitoring
- Applied Machine Learning
- Model Evaluation
- Data Collection
- Data Validation
- Data Synthesis
- Model Training
- Unstructured Data
- Application Deployment
- Data Quality
- Data Maintenance
- Data Preprocessing
Tools you'll learn
Details to know
6 assignments
See how employees at top companies are mastering in-demand skills
There are 3 modules in this course
In this Machine Learning in Production course, you will build intuition about designing a production ML system end-to-end: project scoping, data needs, modeling strategies, and deployment patterns and technologies. You will learn strategies for addressing common challenges in production like establishing a model baseline, addressing concept drift, and performing error analysis. You’ll follow a framework for developing, deploying, and continuously improving a productionized ML application.
Understanding machine learning and deep learning concepts is essential, but if you’re looking to build an effective AI career, you need experience preparing your projects for deployment as well. Machine learning engineering for production combines the foundational concepts of machine learning with the skills and best practices of modern software development necessary to successfully deploy and maintain ML systems in real-world environments. Week 1: Overview of the ML Lifecycle and Deployment Week 2: Modeling Challenges and Strategies Week 3: Data Definition and Baseline
This week covers a quick introduction to machine learning production systems focusing on their requirements and challenges. Next, the week focuses on deploying production systems and what is needed to do so robustly while facing constantly changing data.
What's included
8 videos3 readings2 assignments1 app item2 ungraded labs
8 videos•Total 75 minutes
- Welcome•10 minutes
- Steps of an ML Project•4 minutes
- Case study: speech recognition•12 minutes
- Course outline•3 minutes
- Key challenges•14 minutes
- Deployment patterns•12 minutes
- Monitoring•11 minutes
- Pipeline monitoring•10 minutes
3 readings•Total 14 minutes
- Join the DeepLearning.AI Forum to ask questions, get support, or share amazing ideas!•10 minutes
- Week 1 Optional References•3 minutes
- Lecture Notes Week 1•1 minute
2 assignments•Total 20 minutes
- The Machine Learning Project Lifecycle•10 minutes
- Deployment•10 minutes
1 app item•Total 1 minute
- Intake Survey•1 minute
2 ungraded labs•Total 90 minutes
- Deploying a Deep Learning model•30 minutes
- Deploying a deep learning model with Docker and a cloud service (optional)•60 minutes
This week is about model strategies and key challenges in model development. It covers error analysis and strategies to work with different data types. It also addresses how to cope with class imbalance and highly skewed data sets.
What's included
16 videos2 readings2 assignments1 ungraded lab
16 videos•Total 107 minutes
- Modeling overview•3 minutes
- Key challenges•5 minutes
- Why low average error isn't good enough•11 minutes
- Establish a baseline•8 minutes
- Tips for getting started•6 minutes
- Error analysis example•8 minutes
- Prioritizing what to work on•6 minutes
- Skewed datasets•12 minutes
- Performance auditing•8 minutes
- Data-centric AI development•3 minutes
- A useful picture of data augmentation•6 minutes
- Data augmentation•9 minutes
- Can adding data hurt?•6 minutes
- Adding features•9 minutes
- Experiment tracking•5 minutes
- From big data to good data•4 minutes
2 readings•Total 4 minutes
- Week 2 Optional References•3 minutes
- Lecture Notes Week 2•1 minute
2 assignments•Total 20 minutes
- Selecting and Training a Model•10 minutes
- Modeling challenges•10 minutes
1 ungraded lab•Total 60 minutes
- A journey through Data•60 minutes
This week is all about working with different data types and ensuring label consistency for classification problems. This leads to establishing a performance baseline for your model and discussing strategies to improve it given your time and resources constraints. This week also includes the final end-to-end project.
What's included
17 videos5 readings2 assignments2 ungraded labs
17 videos•Total 128 minutes
- Why is data definition hard?•4 minutes
- More label ambiguity examples•9 minutes
- Major types of data problems•11 minutes
- Small data and label consistency•8 minutes
- Improving label consistency•9 minutes
- Human level performance (HLP)•10 minutes
- Raising HLP•8 minutes
- Obtaining data•12 minutes
- Data pipelines•6 minutes
- Meta-data, data provenance and lineage•10 minutes
- Balanced train/dev/test splits•5 minutes
- What is scoping?•3 minutes
- Scoping process•7 minutes
- Diligence on feasibility and value•14 minutes
- Diligence on value•7 minutes
- Milestones and resourcing•3 minutes
- Final project overview•2 minutes
5 readings•Total 14 minutes
- [IMPORTANT] Reminder about end of access to Lab Notebooks•2 minutes
- Week 3 Optional References•3 minutes
- Lecture Notes Week 3•1 minute
- References•5 minutes
- Acknowledgments•3 minutes
2 assignments•Total 30 minutes
- Data Stage of the ML Production Lifecycle•20 minutes
- Scoping (optional)•10 minutes
2 ungraded labs•Total 105 minutes
- Data Labeling•45 minutes
- The Machine Learning Project Lifecycle•60 minutes
Instructor
Offered by
Explore more from Machine Learning
- Status: Free Trial
Course
- Status: Free TrialC
Coursera
Specialization
- Status: Free Trial
Specialization
- Status: Free TrialG
Google Cloud
Course
Why people choose Coursera for their career
Learner reviews
- 5 stars
84.07%
- 4 stars
12.98%
- 3 stars
1.90%
- 2 stars
0.74%
- 1 star
0.29%
Showing 3 of 3363
Reviewed on Jun 4, 2021
really a great course. It'll really change your way of thinking ML in production use and will help you better understand how can you leverage the power of ML in a way that I'll really create a value
Reviewed on May 19, 2021
Excellent course, as always! Many thanks! Great combination of theory + notebooks with practical examples.Everything is perfectly structured. I will recommend this course to everyone!
Reviewed on Jan 7, 2023
I really enjoy participating in a great class like Andrew's class. It's full of useful and applicable points that I encounter during a real prj. Thanks for sharing this asset with us :))
Frequently asked questions
Machine learning engineering for production refers to the tools, techniques, and practical experiences that transform theoretical ML knowledge into a production-ready skillset.
Effectively deploying machine learning models requires competencies more commonly found in technical fields such as software engineering and DevOps. Machine learning engineering for production combines the foundational concepts of machine learning with the functional expertise of modern software development and engineering roles.
Understanding machine learning and deep learning concepts is essential, but if you’re looking to build an effective AI career, you need production engineering capabilities as well. With machine learning engineering for production, you can turn your knowledge of machine learning into production-ready skills.
The Machine Learning in Production course covers how to conceptualize integrated systems that continuously operate in production as well as solve common challenges unique to the production environment. In striking contrast with standard machine learning modeling, production systems need to handle evolving data. Moreover, the production system must run non-stop at the minimum cost while producing the maximum performance.
In this Machine Learning in Production course, you will build intuition about designing a production ML system end-to-end: project scoping, data needs, modeling strategies, and deployment patterns and technologies. You will learn strategies for addressing common challenges in production like establishing a model baseline, addressing concept drift, and performing error analysis. You’ll follow a framework for developing, deploying, and continuously improving a productionized ML application.
By the end, you will be ready to:
Design an ML production system end-to-end: project scoping, data needs, modeling strategies, and deployment requirements.
Establish a model baseline, address concept drift, and prototype how to develop, deploy, and continuously improve a productionized ML application.
Build data pipelines by gathering, cleaning, and validating datasets.
Implement feature engineering, transformation, and selection with TensorFlow Extended.
Establish data lifecycle by leveraging data lineage and provenance metadata tools and follow data evolution with enterprise data schemas.
Apply techniques to manage modeling resources and best serve offline/online inference requests.
Use analytics to address model fairness, explainability issues, and mitigate bottlenecks.
Deliver deployment pipelines for model serving that require different infrastructures.
Apply best practices and progressive delivery techniques to maintain a continuously operating production system.
More questions
Financial aid available,