How to get started with Machine Learning: Part 1 with Deepesh Wadhwani | Ep #20 | ML for beginners

Are you planning to learn Machine Learning but don't know where to start? Watch Episode 20 to find out how to get started with Machine Learning, where to begin and how to target developing your ML skills. TIME-STAMPED SHOW NOTES: [00:10] Topic of Discussion: How to get started with Machine Learning? [00:30] Series Recap [01:06] Welcome Deepesh! Senior Faculty at Accredian [01:40] How to get started with ML tools? [01:50] Checkpoints to cover to get started [02:53] Objectives and approaches of an ideal ML classroom [04:53] How much math goes into ML? [05:54] Wrap Up! [06:00] Learn more about Data Science at www.accredian.com Accredian is India's first and one of the largest academic institutions dedicated to research and education in AI, Data Science, Product Management, Business Management, etc. Visit us at https://www.accredian.com

What does the episode recommend as starting checkpoints?

The conversation lays out a practical order: get comfortable with Python first, then pick up the core ML tools, then study the math you actually need as it appears. That sequence mirrors Google's Machine Learning Crash Course, which assumes basic Python and introduces tools and concepts together through worked exercises rather than front-loading theory.

How much math do you need to start machine learning?

Less than most beginners fear, and more than zero. To follow introductory material you need three pillars: linear algebra (vectors and matrices), calculus (derivatives and gradients), and basic probability and statistics. The deep end — measure theory, optimization theory — can wait until you specialize. A workable approach is to learn each math concept when a model forces the question: gradients when you meet gradient descent, matrices the first time you shape a dataset.

Which free tools should beginners install first?

• Python 3 — the default language of ML; the official documentation at docs.python.org includes a full beginner tutorial.
• scikit-learn — classical algorithms (regression, trees, clustering) with a uniform API; the getting-started guide runs in minutes.
• Kaggle notebooks — free browser-based environment with datasets included, so no local setup blocks progress.
• A notebook environment — Jupyter or Google Colab for mixing code, plots, and notes.

What should you do after listening?

Pick one small dataset and take a model end to end: load it, train a simple classifier, measure accuracy, and write down what surprised you. Kaggle's free Intro to Machine Learning course does exactly this with housing-price data. Momentum comes from finishing small projects, not from collecting more course recommendations.

Sources

• Google — Machine Learning Crash Course: developers.google.com/machine-learning/crash-course
• Python — Official tutorial and documentation: docs.python.org
• scikit-learn — Getting Started: scikit-learn.org/stable/getting_started.html
• Kaggle — Learn: kaggle.com/learn

Recommended gear for this workflow

• Deep Learning (Adaptive Computation and Machine Learning series) — Structured books beat scattered tutorials when you are building fundamentals.
• Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow — Structured books beat scattered tutorials when you are building fundamentals.
• Pattern Recognition and Machine Learning (Information Science and Statistics) — Structured books beat scattered tutorials when you are building fundamentals.

As an Amazon Associate I earn from qualifying purchases.