Generative AI System Design Interview

by Ali Aminian, Hao Sheng

Genres: Artificial Intelligence

377 pages, Paperback

Published November 16, 2024

Reviews

[Yaron · Rating 4 out of 5]

Seems a bit silly to review an interview book. But I found it solidly written and I learned/solidified a bunch of Generative AI concepts by reading the book. So I'd recommend it even for someone who has a dated machine learning background and just wants to learn about generative AI.

Two caveats is I suspect a lot of the ideas are dated at this point as Gen AI is a fast moving field. And the actual system design component is weak and you'd have to supplement your system design + diagraming skills elsewhere. But obviously interviewing is very heterogenous and no single resource can cover all interview styles.

[Adam · Rating 4 out of 5]

I thought it was very good for the purpose, which is to give the high level overview of relevant techniques and approaches. The right level of detail in a book that aims to cover multiple hot areas in GenAI. If it skipped more,it would have been too shallow. If it included much more, it would have been twice the size. The references in each chapter help the interested reader go deeper.

It could have been somewhat improved by reducing some redundant content between chapters and using that space to go a bit deeper.

[Shruti Pandey · Rating 4 out of 5]

Forgot to add the only technical book I read this year lol. Read this for interview prep in July this year. I used the exact same framework in an interview and got a good review, so it's a testimony that this book works, but I feel like it could have been a booklet instead of being a book (could have cost less lol) because the case studies were very similar to each other.

I love the ML system design interview book - it has a greater variety of case studies and topics that are asked in interviews, so both books in combination are super helpful.

[Mikhail Filatov · Rating 2 out of 5]

The authors took the following approach:
1. Take a paper about some Gen AI features, like text to image
2. Frame it as an interview question
3. Summarize a paper.
4. Rinse and repeat.
There were a couple of chapters, e.g. on RAG, which were more or less relevant for a system design interview question, but majority were absolutely generic and high level.