Where to start for ML engineer ?

I have mentored many mle s.
Do python (proficient enough to solve any leetcode easily) Learn numpy, pandas and matplotlib packages (read the user guide given in the documentation for each and visualise how vectors are being made in numpy, master how slicing and other matrix operation work. No need to remember all the functions) Do kaggle ml intro course and start Kaggle competitions. This is best way to get into ml. Kaggle has active competitions from very basics of tabular data, time series forecasting all the way up to complex NLP, llm, computer vision etc competitions. The competitiveness keeps you motivated to learn more and code more. I recommend doing 3-4 playground series competitions using sklearn and other tabular ml python packages and then move to domain specific competitions. You will learn A TON from public notebooks, and they are very well structured, and you'll be applying concepts from day 1. Continue this for 3-4 months while in parallel learning theory and maths starting from probability stats and going to neural networks, and read research papers (classic - attention is all you need paper). For interviews make sure your theory is strong especially for faang.

DO NOT USE AI chatgpt or anything else when learning!!! Don't waste time on vague YouTube videos, or paid courses. do everything hands on. It's engineering. No need to spend any money, everything i mentioned is free.
I am world top 50 on kaggle, top 5 in india. Won ml competitions worth ~20k usd in my bachelors, and also got into amazon via their ml challenge (rank 1)

You can learn in this order:

Python, Numpy, SciPy, Matplotlib, Seaborn, classic ML algorithms, Scikit-learn, Deep Learning, Tensorflow, Reinforcement learning, GAN, Diffusion models, Computer vision (OpenCV), NLP, LLMs, Agents (LangChain, LangGraph, CrewAI), MCP, A2A, n8n

+1