Skills to build
- Careful reading of annotation guidelines and edge cases
- Spreadsheets, data quality checks and basic Python or SQL
- Language skills, including Indian languages, for text and speech data
- Privacy, consent and confidentiality practices
Roles and responsibilities
- Label and review data to documented guidelines.
- Flag unclear, biased or sensitive examples for review.
- Track quality measures and suggest guideline improvements.
How AI can help
Use pre-labelling tools that suggest labels, correcting them where they are wrong.
Where people take responsibility
People judge context, fairness and sensitive content, and decide when data should not be used.
Suggested preparation
Any degree or diploma with strong reading and digital skills, plus a practice project that documents guidelines, quality checks and corrections.