Machine Learning Interviews, O'Reilly book, behind the scenes process

What's it like writing a technical book with a publisher?

I’d been blogging for a few years on susanshu.com, sharing my experiences in the machine learning field with the goal of helping my readers level up in their ML and data careers.

An editor from O’Reilly (Nicole) reached out for me to teach a bootcamp on O’Reilly on ML interviews. A few weeks later, this coincidentally led to a conversation on a book proposal; since I was already teaching a course on ML interviews, there was a lot of content I could use in tandem for a potential book.

There was some back and forth before and after I submitted my book proposal. There is an approval process with several rounds of O’Reilly editors; even if I had submitted a book proposal, it wasn’t a guarantee that my proposal would be selected. But fast forward a few months later, I got the greenlight to proceed and sign the contract.

Afterward, I got assigned a development editor (Sara), who worked with me to set a schedule, deadlines to hit, and overall also acts like a project manager for me.

So far, I’ve been fairly on schedule, and we use a spreadsheet to track progress. It’s been very fulfilling to see the grid fill up!

After I submit each chapter, the editor reviews them, and sends it to technical reviewers that are leaders in the field (in batches of 2-4 chapters). After the drafts are all complete, there are several more months dedicated to editing, copyediting, and responding to reviewer feedback.

The early access version in my case is truly unedited — there is a copy that is frozen before I go back to make edits. This is optional, as O’Reilly respects each writer’s level of comfort with putting out their unedited content. However, since I’m used to publishing written content online, I eventually decided I was fine with it, even if the preview isn’t the best, in my standards. This is similar to “learning in public”, which I enjoy doing, but had less opportunity to do in the past few years.

There have been challenges, including a busy speaking and travelling schedule: for example I went to Berlin as a keynote speaker at PyCon DE & PyData Berlin, and Orlando and San Francisco.

Occasionally, I face writer’s block. I’ve had many years of practice writing, so this was actually a smaller challenge; I have several personal tactics I’ve used for years to reduce it. Here is a detailed write-up: Link

My tips for writing a book-sized project are:

• Five imperfect sentences are better than Zero

  • Better have some crappy sentences to edit, than nothing at all

• Getting started » anything else

  • Avoid getting paralyzed by wanting to write the “perfect” sentence

On energy management: I don’t do binge sessions anymore, I’d say the longest I write is for 3 hours. Usually, it’s only for 1 hour a day. Slow and steady!

Check out the book preview here: Link (Requires O’Reilly online subscription. It’s possible your local library has access for free, for example, TPL has it.)

I’m very excited for this book to come out, and to help folks interested in the ML field. I’m also designing another course on Machine Learning Careers, here’s your chance to sign up for early access: Link

P.S. I’ve gotten my cover animal! The draft is still a work in progress and I’ll share it once some more changes are made 😄