Machine-learning Engineer Growing path

ML learning path

By difficulties in general :

Intro:

1. Regression(Linear, logistic)
2. KNN, K-means, desecion tree
3. Artifical neuron network(MLP/CNN/LSTM)
   1. Backpropagation, Forward Propergation
   2. Max pooling, Activation Functions
4. RL: bandit/MDP/DP/TD learning/MC
5. Coding:
   1. Python (Numpy, matplotlib, scikit-learn, open-cv , pytorch, Tensorflow etc)
   2. Good to have: GPU programming (CUDA C++ etc)
   3. Good to have: Parallel Computing
   4. Good to have: Matlab, R, Julia
   5. Gits, Unix enviorment

Beginner:

1. Bayes optimization, Bayesian Linear Regression/logistic Regression
2. FFT
3. NLP Transformer/BERT/GPT
4. GMM/HMM
5. EM/Baum-Welch/Viterbi/KalmanFilter(UKF)
6. Graphical model/Belief propagation
7. Junction tree
8. SVM/Kernel Method/RKHS
9. model based RL/Policy gradient/actor-critic

Medium: Missing Cuz i am still learning

Hard: Missing Cuz i am far away , hhh

Math you need:

Intro: Calculus, Linear Algebra, probability theory

Beginner: Convex optimization(Ex. Constrained opt/kkt condition）, Graph theory, High-dim probabilty theory(Ex. Normal distribution and its linear transform, conditional probablity), Bayesian inference， Stochastic process

中级：概率论（数族分布，GLM，测度论，skewness，kurtosis），凸优化，统计力学（spin glass model），泛函分析

cite source

Effiency way to manage your code and conduct your experiement

Deep learning Enterprise Intergration

1. code package
2. Rule to create file & directory
3. Management software

https://github.com/youwei1-sudo/MachineLearning-Study-Path/wiki/Enterprise-Integration(代码管理)

Developer Knowledge / learning path(backend)

https://roadmap.sh/backend

General Deep Learning tips

https://jeffmacaluso.github.io/post/DeepLearningRulesOfThumb/

Fine Tuning guide

https://github.com/google-research/tuning_playbook

ML course i think good for ML basis

Machine Learning

10-601, Spring 2015

Carnegie Mellon University

Tom Mitchell and Maria-Florina Balcan http://www.cs.cmu.edu/~ninamf/courses/601sp15/lectures.shtml