I’m not saying you need to be an expert in advanced calculus to do machine learning…
BUT, there is a big difference between someone that does vs someone that does NOT have a good foundation in stats when it comes to getting & explaining business results.
My thought process back in the day was to obtain a great foundation in stats and machine learning at the same time.
So here’s what helped me. I read a ton of books.
Here are the 3 books that helped me learn data science the most...
1. R for Data Science (Wickham & Grolemund) r4ds.had.co.nz
A group of 50 AI researchers (ByteDance, Alibaba, Tencent + universities) just dropped a 303 page field guide on code models + coding agents.
And the takeaways are not what most people assume.
Here are the highlights I’m thinking about (as someone who lives in Python + agents):
1) Small models can punch way above their weight
If you do RL the right way (RLVR / verifiable rewards), a smaller open model can close the gap with the giants on reasoning-style coding tasks.
2) Python is weirdly hard for models
Mixing languages in pretraining helps… until it doesn’t. Python’s dynamic typing can create negative transfer vs. statically typed languages. Meanwhile pairs like Java↔C# or JS↔TS have strong “synergy.”