My Honest Advice to Beginner ML Students for 2025

Summary

In this video, the speaker shares valuable advice on how to learn machine learning (ML) and become an ML engineer or data scientist. Drawing from their six-year journey, they emphasize the importance of mastering the basics, such as classical ML models and fundamental concepts, before diving into more advanced topics like generative AI. They advocate for hands-on, project-based learning and stress the need for strong software engineering skills. The speaker highlights that success in ML takes time, persistence, and a genuine passion for the field, offering encouragement for those starting their ML journey.

Takeaways

• 😀 Focus on mastering foundational ML models like linear regression, SVMs, and random forests before jumping into newer trends like generative AI.
• 😀 Classical ML techniques and concepts (e.g., precision, recall, confusion matrices) are crucial for interviews and meaningful ML discussions.
• 😀 Start working on real ML projects as soon as possible to gain practical experience, as ML is an empirical and exploratory field.
• 😀 Don't obsess over finding the perfect next course—learning by doing is the most effective way to improve in ML.
• 😀 Software engineering skills are just as important as ML knowledge, especially for debugging, building pipelines, and improving efficiency in projects.
• 😀 Don’t get overwhelmed by new libraries or tools. Stick with foundational tools (like Pandas) and learn newer ones when necessary.
• 😀 It’s more important to understand ML techniques and how they work in practice than to read every new paper or follow every new tool release.
• 😀 Efficiency techniques like Flash Attention and retrieval-augmented generation (RAG) are critical for scaling ML models and systems in production.
• 😀 Be patient—becoming proficient in ML takes time, effort, and consistent practice. Expect to invest thousands of hours into learning.
• 😀 Following the 10,000-hour rule: Mastery comes with sustained practice and deep involvement in real-world projects, not just completing courses.
• 😀 Enjoy the learning process and embrace setbacks. ML is a journey of constant growth, and perseverance will lead to success.

Q & A

What is the most important piece of advice for someone starting to learn machine learning?

The most important advice is to first focus on mastering the basics of machine learning, such as classical models like linear regression, SVMs, and random forests, as well as understanding key metrics like precision, recall, and F1 score. This foundational knowledge is essential before diving into more complex areas like generative AI.

Why is it important to learn the basics of classical ML models before moving on to advanced topics?

Classical ML models form the core of machine learning knowledge. Understanding these models is crucial for interviews and real-world applications. Additionally, having a solid grasp of these basics allows you to engage in meaningful discussions with other ML engineers, helping you build credibility in the field.

What role does math and programming play in learning machine learning?

Math (especially linear algebra and calculus) and programming are fundamental prerequisites for machine learning. A solid understanding of these areas is necessary to grasp ML algorithms, work with data, and implement solutions effectively. Without this foundation, it’s difficult to progress in more advanced ML topics.

How should someone approach their learning path after covering the basics of ML?

After learning the basics, it’s crucial to start working on real ML projects. The field of ML is empirical and exploratory, meaning practical experience is key. You will learn by doing, experimenting, and solving problems in actual projects, even if it feels overwhelming at times.

What is the benefit of focusing on building real-world projects rather than just taking more courses?

Building real-world projects allows you to gain hands-on experience and develop practical skills. Projects expose you to the challenges of debugging, data processing, and building complete systems, which are essential parts of working in ML. This experience also improves your ability to demonstrate your skills through a portfolio.

What should be the focus when creating ML projects for learning purposes?

The focus should be on learning and improving your skills, not just creating impressive projects for your resume. Through iterations, you will see how much you’ve improved with each project, making your work more complex, interesting, and ultimately more valuable to potential employers.

How important are software engineering skills in the field of ML?

Software engineering skills are crucial for ML engineers and researchers. Many ML tasks require solid programming abilities, debugging skills, and efficient code. The process of developing scalable and maintainable systems in ML relies heavily on good software engineering practices.

Is it necessary to keep up with every new ML library or tool?

No, it’s not necessary to keep up with every new library. While learning new tools can be beneficial, beginners should focus on mastering core libraries like Pandas and only gradually explore other tools as their projects evolve. It’s important not to get distracted by the latest trends at the expense of building a strong foundation.

What is the key takeaway regarding the speed of progress in machine learning?

The key takeaway is that learning machine learning takes time. Expect to invest significant effort over several years to become proficient. Don’t be discouraged by the speed at which the field evolves or by the idea that expertise can be achieved quickly. Focus on consistent practice and learning through projects.

How long does it take to become proficient in ML, and what factors influence this timeline?

The time it takes to become proficient in ML depends on factors such as your starting point, prior experience in programming and math, and how much time you dedicate to learning. While the 10,000-hour rule is often cited for mastery, consistent practice over a period of years is necessary to develop real expertise.

What advice does the speaker offer for staying motivated during the learning process?

The speaker advises staying curious, enjoying the process, and pushing through challenges. Machine learning can be tough, but if you enjoy it and commit to learning from each project, the long journey will eventually lead to great rewards, just as it did for the speaker after six years of dedication.