Writer Highlight: Christoph Molnar
Christoph writes about advanced machine learning topics, including modelling paradigms and interpretability issues.
Welcome to another Writer Highlight, the section where we interview both new and seasoned technical writers. This time, we interview from , a machine learning expert and online educator whom I first met via his book Interpretable Machine Learning. I’m delighted to host him today in our community and share his work with you.
First, tell us a bit about yourself. Who are you, what is your day job, what are your professional interests?
Hey there! My name is Christoph Molnar, and I write technical books on machine learning for a living.
After I finished my PhD in 2022, I struggled to figure out what to do next. Started a postdoc. Quit after 3 months. Started in industry. Quit after 3 months. I had lost my love for academia and didn’t feel like going into a corporate job. To figure out my next step, I looked for the things that gave me consistent joy.
This turned out to be technical writing.
I gathered my courage and made the leap to becoming a full-time writer. That was 1.5 years ago, in 2022. Today, I'm happy to say that it worked out and I'm really enjoying my new career.
The topics I write about are guided by the intersection of my own interests and what resonates with my readers. My interests are mostly in machine learning, but since I originally studied statistics, I focus on things that are “missing” in machine learning. Topics I have covered so far include interpretable machine learning, the different ways of thinking about modeling data and quantifying uncertainty with conformal prediction.
Now tell us about
. What is it about? What topics do you cover? Who is your intended audience?Mindful Modeler is about advanced modeling topics in machine learning, such as machine learning interpretability, conformal prediction, and ways to see machine learning. Since my focus is on writing books, Mindful Modeler also serves as a newsletter for my book announcements, and I also share insights about my writing process. My readers are machine learning experts and statisticians.
How, when, and why did you begin writing about technical topics on the Internet?
Between my Master's and my Ph.D., I worked in the industry for a few years. I had an 80% position and used that free day a week to learn and work on side projects. A topic I was drawn to was interpretable machine learning. I started reading papers. And I realized that there wasn’t much online content on this topic. So I started to write an open book about how to interpret machine learning models, which was and is available for free online. This was actually quite weird because I hated writing my Master’s thesis. But it turned out I don’t hate writing itself. I just prefer writing on my own terms and in my own format.
Before writing the book, I started a couple of blogs that fizzled out. But looking back, I realized these were the first foundation stones for my writing career.
What role does technical writing play in your job and life?
For 6 years technical writing was just a side project, but now technical writing plays a big role: it’s my daily job and my main source of income. I’ve published 4 books, write a weekly newsletter named Mindful Modeler for almost 8k subscribers, and publish regularly on Twitter and LinkedIn. I’m also working on a new book Supervised Machine Learning for Science, together with my former colleague Timo Freiesleben.
How do you organize your writing schedule? How much time does writing take you, on average?
Since I'm a full-time writer, I can plan my schedule around writing. I have found that I can only write and edit for a few hours a day, maybe 3 to 5. So, I usually start at 8 in the morning and do most of my writing before lunch. But I also have bursts of writing in the afternoon. Since I only self-publish, I have a lot of other things to do to support my writing, such as marketing, organizing reviews, posting on social media, formatting, and so on.
How is your writing process?
My writing process itself is evolving all the time. Most books I’ve written so far had a simple structure where each chapter was more or less self-contained. This made writing easier because it allowed me to tackle each chapter individually. Once I had the rough outline of a book, I could take one of the chapters and work through it. My rough approach looks like this:
Dump all code examples, research results, and thoughts into the chapter file.
Structure the chapter with subtitles.
Write a very fast bad draft.
Do multiple edit rounds.
In reality, of course, the process isn’t that linear, and I often jump back and forth. When I edit the chapter and realize a subsection is missing, I jump right back to doing research, for example.
Two months ago, I discovered note-taking for myself, and I am quite excited about it. This means I’m trying a different approach for writing books, which is to collect extensive notes first, and only very late in the process, I link all the notes together to a shared narrative, which is the book. I have no idea if it will work for me, but I want to give it a try. So far, I enjoy the note-taking part since I’ve become much more systematic with managing references like books and papers, and I feel like my thinking has improved.
What kind of advice could you give people considering or getting started with technical writing?
Separate writing and editing: When you sit down, decide whether you are in writing or editing mode.
In writing mode, your goal is to create a lot of text, and you aren’t allowed to overthink sentence structures, word choices, and so on. Just write.
In editing mode, your goal is to take the writing that you have and improve it.
Again, it’s not always simple to completely separate these steps, but especially allowing myself to be in writing mode without having to worry about the text quality helped me overcome writer's block.
Don’t be too attached to your words. Cut and throw away generously when you are in editing mode. Cutting the clutter is essential to writing clearly.
Create small publishable packages with your copy. Think posts for social media, a newsletter, blog posts, etc. This will help you hone your skills, take the fear out of publishing, and create tight feedback loops.
Any closing words you’d like to share with the readers of The Tech Writers Stack?
When ChatGPT came on the scene, I doubted my career choice as a writer. My fear was that writing could be fully automated. And it has come true! The web is now full of AI-generated content, and it's only going to get worse. You have already lost – but only if your goal is to write boilerplate content that sounds plausible.
Your readers aren't just interested in how good-sounding your content is. They read your texts because of your unique insights, writing style, and personal story. And while AI-generated text may add more noise to the Internet, it's a chance for writers to stand out with a strong signal.
Huge thanks to Christoph for sharing his story and his advice with our community. Subscribe to if you haven’t, it’s truly a treasure trove for any machine learning practitioner or enthusiast.
Are you a tech writer yourself? Come share your story with us!
 | A guest post by
|
This was a great read. Many points that I resonate with.
This one deserves more attention:
"Don’t be too attached to your words. Cut and throw away generously when you are in editing mode. Cutting the clutter is essential to writing clearly."
I always struggle with thinking I was just SO clever with those words I picked, so that makes it tough to get rid of them... but this is the way!