For those interested in other resources, this is a running list blogs, articles or talks that have either been foundational in how I think about data science and software or something that I routinely visit to stay in touch with where the industry is going and gain perspective on how others think about machine learning.
I’ll try and keep this list short and relevant.
If I had to recommend just one resource to data scientists who are just getting started it would be Google’s machine learning guide.
Blogs
I’m probably guilty of visiting Andrew Gelman’s blog more than any other. Although most of the content is highly specific to Bayesian statistics I think his perspective on problem solving is relevant to a lot of machine learning.
In no particular order, here are some other blogs I like to follow up on (although some haven’t posted new content in a while).
Bayesian Inference
Below are some great resources on getting started with Bayesian Inference.
- Jake Vanderplaas has an excellent talk and corresponding blog post on highlighting the differences between statistics traditionally taught in undergraduate courses and Bayesian inference.
- Wikipedia’s Conjugate Pior Cheat Sheet
- Stan has a great reference for choosing priors
Industry leaders
From time to time I check in on what companies like Google, DeepMind, OpenAI and Microsoft are up to.
Talks on Machine learning
One of the best intros to deep learning in my opinion: Sequence to Sequence Deep Learning - Quoc Le.
A talk about Netflix’s Metaflow: Human-Centric Machine Learning Infrastructure @Netflix. I’m still not persuaded to personally adopt Metaflow but this is an excellent overview of the challenges any company will face when trying to grow a data science team.
This talk from PyData NY 2018, Learning in Cycles: Implementing Sustainable Machine Learning Models, gives great guidelines on approaching machine learning applications without getting boxed in and leaving room for iterations.
Talks on Software
On software design and practice Good Enough is Good Enough - Alex Martelli was an invaluable resource.
Anything by David Beazely is guaranteed to be entertaining and teach you something new about Python.
Raymond Hettinger’s talks are great for those new to learning Python.
How to Write Deployment-friendly Applications - Hynek Schlawack contains some great advice for thinking about deployments.
Microservices
The following talks were extremely helpful when I first began thinking about deploying machine learning models behind REST APIs.