← All courses

AI & Machine Learning

Build, train and evaluate machine learning models — and judge honestly whether the result is good enough to use.

advanced14 weeks

Hunarsaaz course. Written by Hunarsaaz. The certificate is issued by Hunarsaaz.

Who this is for
Analysts and developers moving into machine learning, and graduates in computing, engineering, statistics or mathematics.
What you need first
AI & Data Science, or equivalent Python and pandas ability. Comfortable with basic statistics. This is the most technically demanding course in the catalogue.

What this course sets out to do

  • Frame a real problem as a machine learning task, or recognise that it is not one.
  • Prepare data properly for training.
  • Train supervised and unsupervised models.
  • Evaluate a model honestly, using the right measure.
  • Recognise and reduce harm from a deployed model.

What you will be able to do

  • Distinguish supervised, unsupervised and reinforcement learning.
  • Split data into training and test sets, and explain why.
  • Train classification and regression models with scikit-learn.
  • Use clustering to find structure in unlabelled data.
  • Interpret accuracy, precision, recall and a confusion matrix.
  • Explain why an accurate model can still be unusable or unfair.

What you will cover

3 modules · 9 lessons

  • Types of Machine LearningPreviewVideo coming soon20 min
  • Features, Labels and SplitsVideo coming soon24 min
  • Feature EngineeringVideo coming soon24 min
  • Module 1 — Concepts and Data Preparation — Quiz

Where this can lead

  • Machine Learning Engineer (junior)
  • Data Scientist
  • AI Developer
  • Research Assistant
  • Analytics Engineer

These are the roles the skills apply to. Hunarsaaz does not guarantee employment.