This isn’t just a Machine Learning Operations book.There are a lot of gotchas and pitfalls involved with making an LLM work in production
because LLMs just don’t work like traditional software solutions.To turn anLLM into a product that can interact coherently with your users will requirean entire team and diverse set of skills.Depending on your use case you mayneed to train or finetune and then deploy your own model,or you may justneed to access one from a vendor through an API.
Regardless of which LLM you use,if you want to take full advantage of thetechnology and build the best user experience,you will need to understandhow they work.Not just on the math/tech side either,but on the soft side forhow to make it a good experience for your users.In this book we’ll be covering everything you need to make LLMs work in production.We’ll talkabout the best tools and infrastructure,how to maximize their utility with prompt engineering and other best practices like controlling costs.LLMs
could be one step towards a greater equality,so if you are thinking,”I don’tfeel like the person this book is for,”please reconsider.This book is for thewhole team and anyone who will be interacting with LLMs in the future.