Machine Learning Product Manual
free ebook from Hypergolic
Six months ago, my business partner Chris Kelly and I wrote "Machine Learning Product Manual", an ebook that we made available for free (for now). We did this to popularise our LeanML framework, seek feedback on our methods and eventually use it as an instruction manual for those who want to work with us.
I am pleased to report that we updated the book, and version 0.0.5 is available from https://machinelearningproductmanual.com/. If you would like to keep in touch regarding new versions of the book, select "subscribe" in the popup form, this will allow us to email you in the future.
The writing was an epic 4-week rush which resulted in 100+ pages of text and numerous graphics. We needed to resolve numerous technical challenges and overcome niche details like embedding SVGs in PDFs and how to select a license. If you are planning to write your own book, get in touch for pointers.
We used several sources for inspiration for the content, most importantly, Microsoft's Team Data Science Process and our experience at a London fintech startup. Further sources were: CRISP-DM, Agile/Lean methodology, various Business Canvases and many more. We felt a need for a comprehensive resource tying together the non-technical aspects of managing ML products. You will find meeting agenda templates, organisational maps, and iterative processes tailored to resolve the challenge of incremental industrial-scale Machine Learning projects.
If you are pressed for time, there is a “How to read this book?” chapter that will point you to overview summaries of each major part in the order we think will give you the steepest learning curve. You can revisit the book for the details later.