Three full-length practice tests for the Oracle Cloud Infrastructure AI Foundations Associate (1Z0-1122-26) certification exam - 40 questions per test, 60 minutes, 65% to pass, unlimited retakes.
Every question is grounded in Oracle's official exam objectives and product documentation - written fresh for this course, not recycled from public dumps.
What you'll be tested on
- AI, machine learning, and deep learning fundamentals
- OCI AI Services: Language, Vision, Speech, Document Understanding
- OCI Data Science and the machine learning lifecycle
- Generative AI and large language model concepts
- OCI Generative AI service and AI infrastructure (GPUs, superclusters)
Where candidates lose points
Most people preparing for this exam come from a cloud or ops background rather than a data science one, so the machine learning and deep learning foundations sections - model evaluation metrics like precision, recall, and F1, the bias-variance tradeoff, and how backpropagation and gradient descent actually work - tend to be the weakest spots, not the OCI product features. The product side (which OCI AI service - Vision, Speech, Language, or Document Understanding - fits a given use case) is usually more straightforward once you've seen the console, but candidates who skip the underlying ML concepts often get tripped up by scenario questions that require picking the right technique, not just naming a service.
Sample question
A team wants to evaluate a machine learning model that predicts equipment failure, where missing an actual failure (a false negative) is far more costly than a false alarm. Which metric should they prioritize when comparing candidate models?
- A. Accuracy
- B. Recall
- C. R-squared
- D. Precision
Full explanations and 119 more questions like this are in the practice tests.
How long to prepare
Two to three weeks is realistic if you're comfortable with basic OCI navigation and have some prior exposure to AI/ML concepts; if machine learning terminology is entirely new to you, budget closer to four weeks given the 40-question, 65%-to-pass format covers both foundational ML theory and OCI-specific services in roughly equal measure.
