
CourseBeta
Knowledge Engineering for AI
Work beside engineers at six California organizations, from a newspaper to a hydro plant, as they get their data ready for AI. In the last chapter you build a knowledge system for an organization you choose and hand it over.
Enroll for FREECourse Content
- 1.1 Listing archive sources
- 1.2 Monitoring a pipeline
- 1.3 Labeling article text
- 1.4 Defining article schemas
- 1.5 Organizing the archive
- 1.6 Measuring assistant answers
- 2.1 Data source authorization
- 2.2 Profiling sensor readings
- 2.3 Profiling values across two sources
- 2.4 Profiling USGS river readings
- 2.5 Building the plant pipeline
- 2.6 Catching bad data
- 2.7 Recovering a missed night
- 3.1 Profiling citizen reports
- 3.2 Writing quality rules
- 3.3 Applying quality rules
- 3.4 Labeling damage grades
- 3.5 Holding back a clean test set
- 3.6 Making extra training photos
- 3.7 Writing the preparation log
- 4.1 Choosing churn features
- 4.2 Writing the feature schema
- 4.3 Tracing the churn score
- 4.4 Granting least-privilege access
- 5.1 Choosing the knowledge sources
- 5.2 Organizing the knowledge sources
- 5.3 Writing the metadata sheet
- 5.4 Mapping the knowledge
- 5.5 Reading the licenses
- 5.6 Writing the rights register
- 6.1 Chunking the circulars
- 6.2 Embeddings find meaning
- 6.3 Backfilling missed circulars
- 6.4 Building the golden query set
- 6.5 Scoring the golden queries
- 6.6 Changing the index on policy day
- 7.1 What goes into the airline's index
- 7.2 Two winter storm notices
- 7.3 Checking the answers at the counter
- 7.4 Running the baggage-rule change
- 7.5 Runbook for indexer failure
- 8.1 Choosing your organization
- 8.2 Governing your assistant's sources
- 8.3 Building your index
- 8.4 Proving the retrieval
- 8.5 Your assistant answers questions
- 8.6 Handing over your system
Frequently asked questions
You can start Knowledge Engineering for AI with a 7-day free trial. Click "Enroll for FREE" to begin.
About 14 hours, across 47 lessons.
Knowledge Engineering for AI is an intermediate course. Prerequisites: Basic programming. The course teaches data engineering and SQL from the start..
Most lessons run in your browser. The final chapter runs on your own computer with Python 3.10 or newer and about 1 GB of space for the embedding model.


