AI safety & evaluation | Sri AI

Course overview

How to find out whether a system is safe before the public does, including for languages that have no benchmark to test against.

Level: Advanced  ·  Mode: Evenings, online

Who this course is for

Engineers and QA leads responsible for signing off releases.

What you will learn

• Build evaluation sets for low-resource languages
• Red-team a deployed system
• Measure refusal and failure behaviour
• Establish a release gate that has teeth

Syllabus

1. Module 1: Evaluation design
2. Module 2: Building benchmarks where none exist
3. Module 3: Red-teaming methods
4. Module 4: Measuring harm and refusal
5. Module 5: Release gates and sign-off

Final project

Every module ends in hands-on practice, and the course ends with a project you build and present. Your certificate names that project.

Before you start

LLM engineering.