Applied Machine Learning and AI for Engineers: Solve Business Problems That Can't Be Solved Algorithmically

To this day, I recall the frustration of failing a college course even though I worked hard and wanted to understand the material. The professor certainly seemed very intelligent. Thankfully, I passed after retaking the course a semester later. The difference was another professor who had a gift for explaining complicated material in a way that made sense to me. Jeff Prosise also has that gift. Don't let the book's title scare you into thinking that to understand its concepts, you have to be a nerd in a dark room drinking Mountain Dew until the odd hours of the night tinkering with code (Oh wait, that's me!), or, you have to be a data scientist or have degrees in advanced mathematics. It is approachable and understandable to anyone with a logical mind and the ability to use Google to define some terms that pop up here and there. I really think managers and CxOs who can't see the random forests for the decision trees will benefit the most in getting a grip on what this topic is all about. I will say something that may seem outrageous, but I think many of the readers here who like sci-fi could skip installing and running the code examples and come away being able to determine science fiction from fact in this topic. This book is broken into two parts: part 1 builds foundations, defines terms, and covers traditional machine learning (ML) and part 2 delves further into deep learning and building neural networks. The entire book has wonderful, real-world examples. It uses the most popular tools like Scikit-Learn and TensorFlow for building machine learning models. It also uses Python for most code examples with a few things in C# for ML.Net and some client examples. I appreciate the real-world examples like predicting taxicab arrivals or credit fraud that connect with actions people perform day to day. The audio classification example, which uses sound files with a convolutional neural network, is fascinating and creative. The chapters covering facial and object recognition were favorites; I had more than a couple of "aha" moments because Jeff did such a great job building on the basics from the beginning. Have you ever wondered how self-driving cars avoid hitting objects? The chapter on Natural Language Processing interested me because I use Duolingo every day in my study of Portuguese as a third language, and I've wondered how the language processor works. Hey, I have a clue now! The book concludes with a tour of Azure Cognitive Services and a final example that is simple and elegant using the Contoso Travel company so many Microsoft developers are familiar with from demos. Speaking of demos, if you want to follow the examples, Jeff has done a great job of explaining how to set up the environment and even created the Docker container image with everything you need to make it simple. I also learned to use Flask to wrap a Python Model in a web service and call it from a C# client. Way cool! Now I can say I'm busy training my model without HR getting upset... but I digress. Thank you, Jeff, for an excellent book!

Same algorithms, same examples, same datasets. This book is just a reprint of all previous AI/Python books - nothing new. If you remove the words “for Engineers,” nothing will change. The book tries to explain some algorithms but avoids any math, which makes it hard for beginners to understand and boring for those who are already familiar with the algorithms.

This is a great book for both people who never tried to approach AI development and for those who started but did not have a full understanding of some topics or libraries. I loved to see the problems solved with different strategies and methodologies and comparing the results. Also, the insights with the linked articles are precious to go deeper in the math which is well hidden by the python libraries. Even if the latest GPT model are not explicitly covered, this is absolutely unnecessary because once you get what's behind, you'll discover that GPT is just a more complex model base on the same concepts exposed here.

I knew literally nothing about machine learning before reading this book - but I know basic algorithms, high school math, and Python. That was enough for me to easily follow this book and get a great start on my path towards working on ML and AI projects in the future. I give it top marks!

This is a well-produced book with all illustrations and code segments in colour. The author walks you through the exercises in each chapter, but does not explain any code. I was then obliged to read up on the code in other resources, such as ChatGPT. This is not necessarily a bad thing, as arguably I ended up with a better understanding of the code, rather than if I had simply read a description of the code in the book.