2025 How to Learn Artificial Intelligence and Prepare for Career Path as Undergraduate student

Hi, everyone, as a current undergradudate student majoring in Artificial Intelligence. Let me briefly describe what I would do if I went back to the summer after graduating from high school. Since Artificial Intelligence is getting more and more popular application, a lot of STEM background student are preparing their career path towards Artificial Intelligence industry. And I would like to share some of my experience and I wish I could see this earlier when I was still in the freshman years. For more information, you can visit my post on DeepNLP AI Community (http://www.deepnlp.org/community) and AI&Robots Question and Answering (http://www.deepnlp.org/question) and follow my posts.

Table of Content

0. High School Preparation

1. First semester of freshman year

2. Second semester of freshman year

3. First semester of sophomore year

4. Second semester of sophomore year

5. Junior Year of Undergraduate Study

Main Content

0. High School Preparation

0.1 Study advanced mathematics

Learn as much as you can. I previously started learning from scratch with 'Basic 30 Lectures' and found it manageable.

0.2 Basic syntax of C/C++

It's nearly impossible to become an expert in any programming language in a systematic way. I recommend learning the basics of programming, particularly variables, functions, classes, and pointers—these are the frequently used elements that you need to understand. Try writing simple programs, like palindromes and Narcissistic numbers. You’ll feel a great sense of accomplishment the first time you get a program to run. Trust me. For the compiler, I recommend using VSCode. If you search 'vscode configure C++' online, there are plenty of tutorials to help you.

1. First semester of freshman year

1.1 Continue studying advanced mathematics

The purpose of studying advanced math is not just to be able to solve problems, but to understand the formulas and principles of machine learning and deep learning models later on.

1.2 Linear algebra: Study it well! There are many online tutorials available.

1.3 C/C++ programming practice: Go to platforms like Luogu or Leetcode to practice simple foundational problems. This is to help you develop the mindset for writing programs: Why write this way, and how to write it well.

1.4 probability theory: This is also very important! Find tutorials online.

2. Second semester of freshman year

2.1 Data structures

Very important! You can learn it alongside solving problems on Leetcode. (Use C/C++, and at this point, you can learn the STL library in C++ which has many data structures built in).

2.2 Algorithms

Very important! Study it alongside solving problems on Leetcode.

2.3 Python

The most important part of the Artificial Intelligence major. First, learn how to install the Python environment, choose a compiler (VSCode or PyCharm are both fine), and install Python libraries. Then, focus on the basic syntax. With a C++ foundation, learning Python will be easy. Many tutorials are available online, so find one. Don’t forget to practice programming. (It's recommended to install Python 3.6 for better compatibility).

2.4 Extra Challenge

If you have extra time, study discrete mathematics. You don’t need to understand it deeply, just get a general idea (you probably won't have time for deep study). This semester, you can also try participating in algorithm competitions like ACM. Whether you win or not doesn't matter; it’s important to test your aptitude in this area.

3. First semester of sophomore year

Sophomore year is critical! You must make sure to seize this time—whether it’s your GPA or competitions, there will be a huge gap.

3.1 Statistical Learning Methods

3.2 Linux operating system

Learn to set up virtual machines, understand simple Linux commands (use Ubuntu, and don’t use the latest version), and set up Python and C++ environments on Linux.

3.3 Computer graphics

This is a core course for the Artificial Intelligence major, study it well.

3.4 Look for professors to work on research projects

Find a direction you’re interested in. Current AI research topics include: Computer Vision: 3D point cloud completion and processing, 3D reconstruction, supercomputing, image processing, knowledge graphs, etc. NLP: Large Language Models, Generative AI. Get to know these fields and choose one that interests you to follow. (It’s recommended to choose a younger professor since they’re more likely to supervise and motivate you, especially in research. If you don’t have supervision, you may give up after a few months). The purpose of research projects is to learn cutting-edge knowledge and, of course, to publish papers! When you’re qualified for postgraduate recommendation, you’ll understand how much a well-written first-author paper can boost your value and help you get offers from better schools.

3.7 Machine learning: Andrew Ng—both are great options.

4. Second semester of sophomore year

4.1 Continue studying machine learning

You won’t finish it in the winter break, so continue.

4.2 Deep learning Book: Ian GoodFella

4.3 Learn PyTorch: Search for official documentation and tutorials online. Learn how to build a neural network and deploy models. (Research supervisors will definitely require you to do this, as deep learning is everywhere).

5. Junior Year of Undergraduate Study

5.1 Study Computer Vision/Natural Language Processing/Generative AI

Depending on your interests, now you can choose a subfield of AI, such as NLP,CV, Generative Artificial Intelligence, Knowledge Graph, etc. If you like language processing, like automatic translation, machine translation, speech recognition, etc., study Natural Language Processing. If you prefer autonomous driving, image processing, object detection, and recognition, study Computer Vision.

5.2 Participate in AI-related competitions

By this time, you should already be able to participate in AI-related competitions! Of course, I believe by now you would have already participated in other competitions (mathematical modeling, data mining, etc.). However, these competitions are not very valuable! The most valuable competition during your undergraduate years is ACM, nothing else. But if you’re not naturally gifted in problem-solving, I wouldn’t recommend it. Most people can’t think of solutions to those problems, and practicing won’t necessarily raise your upper limit. You can try Kaggle, CCF-type competitions, and CVPR-Workshop (which is very difficult). The main goal is to learn from others—look at how they build models and handle data, and learn their ideas. The point of these competitions is not winning but learning to apply the knowledge you have gained quickly.

5.3 Know your Direction

In your junior year, you should know the direction you want to take. If you follow everything I’ve mentioned and do it meticulously, congratulations! Your programming and practical skills will surpass 95% of your classmates and get better opportunities