AI/ML Roadmap for beginner To Advanced

Linear Algebra

• Khan Academy Linear Algebra (Unit 1): A beginner-friendly introduction to core concepts.
• Calculus by James Stewart (Chapter 12 – Vectors and the Geometry of Space): A solid resource for understanding vectors in 3D space.
• 3Blue1Brown – Essence of Linear Algebra: Engaging visualizations to grasp the intuition behind linear algebra.
• DeepLearing.AI’s Linear Algebra for Machine Learning and Data Science – Available for free on youtube. but dont know how long.

Calculus

• 3Blue1Brown – Essence of Calculus: A fantastic series for building intuition about calculus concepts.
• Khan Academy Calculus 1 
• Khan Academy Calculus 2
• Multivariable Calculus by Khan Academy: Offers excellent visuals and intuitive explanations.
• Practice Problems from James Stewart’s Calculus: Solve problem sets throughout the course to reinforce learning.

I’ve done Calculus 1, 2, and Multivariable Calculus from Professor Leonard, but I found his videos too lengthy as he solves so many problems in his lectures. I saw Khan Academy videos too, somehow I felt they are better in developing intuition as you solve problems along with the videos. For Multivariable Calculus, Khan Academy isn’t even optional, you definitely need to go for it.

Professor Leonard Calculus 1 playlist also has pre-calculus too. Khan Academy has its own pre-calculus course.

Statistics and Probability

• Khan Academy Statistics and Probability: A clear and concise introduction to the basics.
• StatQuest Statistics Fundamentals Playlist: Breaks down complex statistical concepts with clarity.

AI Courses:

• Andrew NG Machine Learning Specialization
• Andrew NG Deep Learning Specialization
• Hands on machine learning with Scikit-learn and TensorFlow 

First, Complete NG Machine Learning Specialization and Complete Part 1 (Machine Learning ) in HOML book.

After Completing them, Start Deep Learning specialization and then Part 2 ( Deep Learning and Neural Network) in HOML book.

You need not to worry if you feel like you are lacking concepts in Linear Algebra or  Probability and Statistics that come in these courses. Andrew explains them from scratch intuitively in ML regard.

Projects 

• Titanic Dataset Prediction – Kaggle: https://www.kaggle.com/c/titanic

• Iris Classification – Kaggle: https://www.kaggle.com/uciml/iris

• MNIST Digit Classification – Kaggle: https://www.kaggle.com/oddrationale/mnist-in-csv

• Project tutorials using Scikit-learn: https://scikit-learn.org/stable/auto_examples/index.html

MLOps 

• Deploy simple ML apps with Streamlit: https://docs.streamlit.io/

• Gradio for ML interfaces: https://gradio.app/

• Version control with Git & GitHub: https://docs.github.com/en/get-started/quickstart

Phase 1: Strengthening the Foundation

Linear Algebra

• MIT 18.06 Linear Algebra by Gilbert Strang: A comprehensive course from a renowned educator. MIT Course Link

• Introduction to Linear Algebra by Gilbert Strang (Book): Complements the course with detailed explanations and exercises. (do a simple search free pdf is available online) 

Calculus

• Continue practicing problems from Calculus by James Stewart to solidify your understanding.

Statistics and Probability

• Harvard Stat 110 by Joseph K. Blitzstein: A rigorous yet accessible course on probability. You will find the practice problems, solutions in this course link 
• Introduction to Probability by Joseph K. Blitzstein (Book): A great companion to the course, offering in-depth insights.

AI/ML Course :

CS224n 2024 – Natural Language Processing 

I felt cs224n youtube playlist is not in order. Kindly, please refer to the course link in stanford website, you can find videos order there.

If you find the cs224n harder, optionally you can do DeepLearning.AI’s (coursera) Natural Language Specialization Course.

Projects 

• Sentiment Analysis – Kaggle: https://www.kaggle.com/c/sentiment-analysis-on-movie-reviews

• SMS Spam Detection – Kaggle: https://www.kaggle.com/uciml/sms-spam-collection-dataset

• Simple Chatbot – Free tutorial: https://www.analyticsvidhya.com/blog/2021/06/build-your-own-chatbot-using-nltk-and-python/

MLOps 

• Deploy NLP models with FastAPI: https://fastapi.tiangolo.com/tutorial/

• Docker for containerization: https://www.docker.com/get-started/

• Automate workflow with GitHub Actions: https://docs.github.com/en/actions

Phase 2: Advancing Your Skills

The courses in this phase build on the foundations from Phase 1, diving deeper into specialized topics.

Matrix Calculus

• Matrix Calculus MIT 18.S096 – A short course on Matrix Calculus

Convex Optimization

• Stanford EE364a by Stephen Boyd: A definitive course on optimization for ML.
• Convex Optimization by Stephen Boyd : A comprehensive resource to accompany the course. Reference Book link

Discrete Mathematics

• MIT 6.042J Mathematics for Computer Science: Covers essential topics like combinatorics and graph theory.

Information Theory

• MIT 6.450 Principles of Digital Communication: Explores the mathematical foundations of information theory.

Mathematics for Machine Learning (Free PDF): A concise book that ties together key mathematical concepts for ML.

Projects 

• Recommendation System – Kaggle: https://www.kaggle.com/grouplens/movielens-100k-dataset

• Multi-class Image Classification – Kaggle: https://www.kaggle.com/c/digit-recognizer

• Text Summarization – Kaggle: https://www.kaggle.com/ibtesama/getting-started-with-text-summarization

• Tutorials: Hugging Face Transformers for text tasks: https://huggingface.co/docs/transformers/index

MLOps 

• ML Pipelines with MLflow: https://mlflow.org/

• Deploy models on free-tier cloud: Google Colab, Render.com, Hugging Face Spaces

• Automate data ingestion, retraining, monitoring pipelines

Phase 3: Advanced ML Courses

These machine learning courses emphasize mathematical rigor, making them ideal for advanced learners.

• CS229 by Stanford: A foundational ML course with a strong mathematical focus. Reference Book : CS229 Book
• Deep Learning Book : Bible for Deep Learning
• Pattern Recognition and Machine Learning by Christopher Bishop: A PhD-level text that dives deep into the mathematics of ML.

Also, through out the course, I widely used these two resources for intuition :

• 2D graphing – Geogebra 2D graphing
• 3D graphing – math3d.org

Disclaimer:  This is my current roadmap which I have been following and I will follow in the coming days. I have listed these resources based on their value in my own preparation. I may update or replace them as better options emerge. Since everyone learns differently, feel free to explore other resources that suit your needs. Wishing you the best on your AI/ML journey! 

Disclaimer2 : Almost all of the above resources  I have shared above are publicly available just with a simple search. If you feel like I have violated Copy rights, Kindly contact me. I will remove the links.