15 to 30 random questions in 15 to 30 minutes, by the level you choose: tokens, sampling, prompts, rag, embeddings, agents, evals, safety. After you submit you get your score, your strong and weak topics, and the right answer with an explanation for every question.
Want to revise first? Read the Generative AI and Prompt Engineering interview questions and answers.
The test covers 13 topics across three levels. Each attempt picks 15 to 30 questions at random, by the level you choose, so every attempt is different.
One question from each level, the way they appear in the test. A new set comes on the 1st of every month. Try it, then open the answer.
Answer: Basing the answer on supplied source material, such as documents in the prompt. A grounded answer is tied to sources the system provides, so each claim can be traced and checked. It is the main defence against made-up facts, and it lets the answer use information the model was never trained on.
a = [0.6, 0.8] # length 1
b = [0.8, 0.6] # length 1 Answer: 0.96. For unit-length vectors cosine similarity equals the plain dot product: 0.6 x 0.8 + 0.8 x 0.6 = 0.96. This is why many systems normalise embeddings once and then use the cheaper dot product.
Answer: A, about 0.48. Temperature divides every logit by the same positive number, so the order never changes and A stays on top. At T = 2 its probability falls from 0.59 to 0.48: the distribution flattens, it does not flip.
Take the full Generative AI and Prompt Engineering mock test: 15 to 30 questions and 15 to 30 minutes, by level.
A question counts as right only when you pick exactly the right option, or all the right options where it says so; a question left blank counts as wrong. Your score is the share of the test's questions you got right (15 to 30, by level). The result also scores every topic, so you know what to revise before the interview.
Yes. The Generative AI and Prompt Engineering mock test is free. You need a free CoreHRx Jobs account to start it, so your score and the topics to work on are saved.
It depends on the level: Fresher has 15 questions in 15 minutes, Intermediate has 20 questions in 20 minutes, Experienced has 30 questions in 30 minutes. They are multiple-choice questions picked at random from a larger bank of Generative AI and Prompt Engineering questions at your level, so every attempt is different.
The questions cover Embeddings and Vector Search, Evaluation, Prompt Patterns, Retrieval-Augmented Generation, Safety and Prompt Injection, Cost and Latency, Hallucination and Grounding, LLM Basics, Tokens and Context Windows, Tool Use and Agents, Sampling and Temperature, Structured Output, Fine-tuning vs RAG vs Prompting. Your result shows how you did in each topic.
Fresher if you are a student or have less than a year of experience, Intermediate for 1 to 5 years at work, and Experienced for 5 years or more.
Only your best Generative AI and Prompt Engineering score, and only when it is 60% or more. It shows on the job applications you send on CoreHRx. Lower scores stay private.
Yes, up to 10 tests a day. Each attempt picks questions you have not seen recently first.