Align possibilities with business objectives. Find Machine Learning use cases that maximize impact & feasibility.

• Introduction to ML

  • What is Machine Learning?

  • No-code ML demos & example use cases

  • Concepts and terminology

• Possibilities and limitations of ML models

  • Types of use cases that create great value

  • Detailed use cases

  • What is Deep Learning?

  • When ML fails

• ML value framework

  • Value Propositions powered by ML

  • Connecting predictions to value propositions

  • The Prediction Task Canvas

  • Monitoring value in production

  • [New] Process Mining: finding ML opportunities in your processes

  • [New] Strategies for delivering ML solutions to your users’ doorstep

Master the Machine Learning Canvas and turn your ML solution idea into a detailed plan. This course will teach you the techniques that innovation teams have used to change the way they think about ML project planning.

Introduction to the Machine Learning Canvas

• Why use the MLC?

• Overview of the 10 boxes that make up the MLC

• Structure of the MLC

• Value vs cost of ML solutions: anticipating costs with the MLC

Listing requirements with the MLC

• Data for ML: features, sources, and collection

• Impact simulation

• Building models and making predictions

Towards a flawless design

• 4 common mistakes with the MLC

• Dealing with feedback loops

• What a bullet-proof MLC looks like (+ case studies)

• Characteristics of a great MLC

• Hands-on (remote) collaboration tips: cloud document vs virtual whiteboard

Avoid pitfalls and prioritize work efficiently with the OWNML Checklist. This tells you, in detail, what you need to do and in which order. End-to-end ML projects are broken down into 9 phases of 4 tasks each. Lessons:

• ML pipelines: understanding the structure of data pipelines that create ML models.

• Software components that need to be built when creating an ML system: architecture diagram + explanations.

• The investor’s approach to ML projects: principles behind the checklist, used to prove value early while minimizing risks and costs of ML projects.

• How to use the checklist + detailed explanations of what it contains.

• Workflow of a model builder.

Our methodology to implement a Minimum Viable Product with no-code/low-code and minimize time to value.

DATA: create your training dataset

• Common data aggregation and transformation operations

• Create feature sets and label sets while preventing data leakage

• Prepare data for ML with a no-code data wrangler

• [New] Low-code data wrangling in Python/SQL with notebooks and GenAI

BASELINE: create your own models in record time with AutoML

• Build classification and regression models with no code

• How to split data into training/validation/test sets while maximizing performance & test reliability

• How to deal with time in ML

• Analyze predictions and errors

• Select the best model

• [New] Low-code modeling with notebooks and GenAI

MVM: turn your baseline into a Minimum Viable Model you can trust

• Explain predictions, understand errors, and improve data prep

• Parametrize & optimize data prep

• Analyze model behavior & performance: from accuracy to business metrics

• Optimize modeling & decisions with practical tweaks

MVP: turn your MVM into a Product / Solution

• Apply your MVM to new inputs

• Automate predictions

• Batch vs real-time predictions

• Deploy to production

• Deliver predictions & recommended decisions with simple end-user interfaces

• Build a dashboard to monitor prediction accuracy and KPIs

• Automate model retraining at the right frequency

• Implement feedback loops

• Tips from a Product Manager to get traction and grow usage