Data Labelling and Annotation Quality
A model is a compressed version of its labels, so careless annotation guarantees a careless model. This course covers running annotation work that produces data worth training on.
Course outline
What this programme covers, module by module.
- 01Models as a compressed version of their labels
- 02Writing guidelines that remove ambiguity
- 03Selecting, training and calibrating annotators
- 04Inter annotator agreement and measuring it
- 05Disagreement, adjudication and edge cases
- 06Ongoing quality audit and label drift
- 07Balancing cost, throughput and quality
What you will gain
- Write annotation guidelines that remove ambiguity
- Select and train annotators for the task
- Measure agreement between annotators
- Resolve disagreement and adjudicate edge cases
- Audit label quality on an ongoing basis
- Manage cost, throughput and quality together
Who should attend
- Data scientists and machine learning engineers
- Annotation and operations teams
- Project managers on data projects
- Quality staff supporting data work
Upcoming sessions
| Date | City | Duration | Fees | Book |
|---|---|---|---|---|
| 2 to 4 Nov 2026 | Durban | 3 Days | R13,000 | |
| 16 to 18 Nov 2026 | Cape Town | 3 Days | R13,000 | |
| 7 to 9 Dec 2026 | Johannesburg | 3 Days | R13,000 | |
| 11 to 13 Jan 2027 | London | 3 Days | US$1,500 | |
| 9 to 11 Feb 2027 | Dubai | 3 Days | US$1,500 | |
| 12 to 14 Apr 2027 | Pretoria | 3 Days | R13,000 | |
| 28 to 30 Jun 2027 | Durban | 3 Days | R13,000 |
Fees are per delegate and exclude VAT. Public intakes are confirmed once minimum numbers are met. We also deliver this course on your own schedule, in-house or live online.
Enquire about this course
Register your team or ask for a tailored in-house quote. We reply within one working day.
WhatsApp usCourse brochure (PDF)- Fees from
- R13,000 per delegate
- Duration
- 3 Days
- Field
- Artificial Intelligence & Emerging Technology
- Formats
- In-house, public, online
More in Artificial Intelligence & Emerging Technology
Artificial Intelligence for Business Leaders
Executives are asked to approve artificial intelligence investments without a technical background. This course gives decision makers enough working understanding to tell a genuine opportunity from a costly distraction, and to ask suppliers the right questions.
Deploying and Operating Machine Learning Models
A model that works in a notebook is not yet a working system. This course covers what it takes to run a model in production: packaging, serving, monitoring for drift and retraining without breaking what depends on it.
Generative AI Tools in the Workplace
Generative tools are already in use across most organisations, often without guidance. This practical course shows staff how to use them well for real work tasks, and where the accuracy and confidentiality traps lie.