After a month off from releasing original interviews on the podcast feed, I’m so excited to be sharing this episode with all of you! Aether Biomachines is one of the most interesting machine learning startups I’ve ever come across and I was thrilled to interview the founder, Pavle Jeremic.

Building a Post-Scarcity Future using Machine Learning

​

In this episode, Pavle talks about his vision for a post-scarcity future, and how his company is using machine learning to accelerate towards that by creating and modeling enzymatic reactions.

​Click here to listen to the episode, or find it in your podcast player of choice: https://www.mlengineered.com/listen​

I’ve started to write out extensive show notes with all the most interesting takeaways and quotes in my personal blog. So if you prefer to read rather than listen, click here to view the post.

And if you don’t have time for either of the above, I’ve distilled the episode even further into this twitter thread. Check it out and please like / RT if you found it helpful!

​

Interest in MLOps is "exploding"?

I subscribe to a few newsletters that keep me updated on various emerging trends in the business world. Machine learning is still relatively niche, so I was quite surprised to see Exploding Topics identify “mlops” as exploding!

​

​

I clicked through to their website and was especially interested to see that “machine learning engineer” was also marked as “exploding”:

​

And while both “machine learning” and “deep learning” have rapidly grown over the past few years and still get more raw searches than the two above, it seems that interest has waned since 2018-19.

​

Quite fascinating to have some (semi-) quantitative information to back up what I’ve suspected was the case: that more and more of the focus around ML has started to shift from solely the research and towards using it to drive real business value.

You can click here to explore more.

​

Machine Learning: The Great Stagnation?

​

A hilarious, interesting, and pointed critique of machine learning by Mark Saroufim blew up on HackerNews last week. While I don’t agree with the magnitude of the critique, I think it’s directionally accurate, especially the sections on “Graduate Student Descent” and “Fake Rigor”.

In the second section, he details a few areas where he thinks there is still substantial innovation occurring including Keras and fast.ai, 🤗, and more. It’s definitely worth a read: Machine Learning: The Great Stagnation​

​

Completing the Machine Learning Loop

​

This medium piece by Jimmy Whitaker really crystalizes the problem we face as ML engineers: that many data science tools and processes that exist treat model development as sequential instead of iterative. This is similar to the way that software was viewed a few years ago until DevOps took a lesson from manufacturing and started to develop best practices that were then solidified into the platforms and frameworks we use today. With this in mind, it’s no wonder that MLOps is such a hot topic (see above).

While the last section reads like an ad for Pachyderm, I greatly enjoyed the article as a whole and recommend it, especially for those who might be new to the topic. And for those more advanced, there are plenty of links to sources I haven’t encountered before where you can dive deeper.

Check it out here: Completing the Machine Learning Loop​

​

A Free Virtual Data Conference

I just heard about this event and think it’s gotten criminally-low attention for what it is: a free conference spanning four Fridays in February, each with a different topic of focus.

It’s hosted by DataTalks.Club and starts Feb 5 with talks on Machine Learning Use Cases.

You can register here: DataTalks.Club Conference​