Generative AI Developer
Build and operate generative AI applications with retrieval, evaluation, safety controls and production observability.
Who this path is for
Know application development, AWS security and generative AI concepts. Build practical experience before this advanced path.
Study each topic, complete the applied exercise, and keep evidence of what you learned. These are original study milestones, not a claim to reproduce the full exam blueprint. Use the linked AWS exam guide to check all in-scope domains before booking.
Your learning roadmap
01Model and knowledge integration
3 focus areas · applied exercise
Model and knowledge integration
3 focus areas · applied exercise- Amazon Bedrock
- Embeddings and retrieval
- Data preparation
Design a grounded document assistant with source citations.
Define how retrieval quality is measured.
02Build the application
3 focus areas · applied exercise
Build the application
3 focus areas · applied exercise- Prompt orchestration
- Tools and agents
- API integration
Design tool permissions and approval boundaries for an assistant.
Explain which actions require human review.
03Evaluate and protect
3 focus areas · applied exercise
Evaluate and protect
3 focus areas · applied exercise- Test datasets
- Prompt-injection defenses
- Responsible AI
Create synthetic adversarial and normal evaluation cases.
Specify groundedness, safety and task-success criteria.
04Operate and optimize
3 focus areas · applied exercise
Operate and optimize
3 focus areas · applied exercise- Token and latency budgets
- Tracing
- Release evaluation
Plan a model upgrade with offline evaluation and monitored rollout.
Define a rollback threshold for quality or cost regressions.
Put the learning into practice
Dedicated CloudAdhar guided labs for this certification are not published yet. Continue with the official AWS preparation resources for additional practice.
Related labs cover selected skills. Completing them does not establish exam readiness.