The Step-by-Step Guide to Becoming a Machine Learning Engineer
And other practical guides to understand machine learning
I get a lot of questions about how to become a machine learning engineer. Everyone asks what they need to know and how they can learn it. I’ve been asked this question so many times, I’ve written it down for everyone to use! Here’s my 100% free guide to becoming a machine learning engineer (and 4 other ways to learn about machine learning): Machine Learning For Everyone.
Machine Learning For Everyone is a repository on GitHub to make it easily accessible to anyone with an internet connection, practical to keep continually updated, and easy to include machine learning practice examples for anyone wanting to train their own model (coming soon). It’s a way for consumers, engineers, and techies to learn whatever they want about ML.
Currently, the guide contains 5 distinct learning paths, which I'll touch on here:
A path to gain the skills necessary to become a machine learning engineer.
A path for anyone interested in AI research who wants to learn about machine learning models.
A path for developers who want to take advantage of the machine learning tools at their disposal in the applications they build.
A path for consumers wanting to understand how machine learning will affect them.
A path for companies who want to use machine learning in their business.
I'm going to touch on each of these paths: who they're for, what they consist of, and where you can find them. If you’re interested in learning more about machine learning engineering (or machine learning in general), join Society’s Backend to learn more about it each week:
Machine Learning Engineer Road Map
This road map is for anyone interested in solving the problems that come with bringing AI to consumers in a practical and productive way. Think about training, experimenting, and deploying these models. Also think about ensuring privacy, quality, and a fantastic user experience.
Machine learning engineers focus on the software engineering side of machine learning and solve problems related to making AI research usable. Machine learning engineers can end up working primarily in software engineering (in a ML application) and might find themselves modeling as well. They need a strong software engineering background, but they also need a good understanding of ML math.
Machine learning engineering is a booming field that has taken off recently as more machine learning applications are brought to the general public. As time goes on, this role will only be more in demand.
This road map takes you through the following prerequisites:
Data Structures
Algorithms
Programmatic Problem Solving
Writing Clean Code
Testing
Version Control
Code Review
System Design/Software Architecture
Linear Algebra
Probability
Python
Lower-level programming languages
It then runs you through Google's Machine Learning Crash Course which teaches a lot of the same principles as CS229 but with more of an engineering perspective. As a machine learning engineer, I also suggest going through CS229 to dig into ML math. Check out the Machine Learning Engineer Road Map here.
More high-quality resources for machine learning engineers will be coming in the future.
Modeling Road Map
The modeling road map is for those who really want to understand the math behind machine learning models. If you really want to answer the question "how can machines possibly learn?", then this is the road map for you.
The modeling road map is geared toward learning the skills necessary to construct and optimize machine learning models. These skills are needed to get in AI research and train large, production-scale models.
This road map starts with the following prerequisites:
Programming
Data Structures
Algorithms
Python
Linear Algebra
Probability
Version Control
It walks you through Stanford's CS229 Intro to Machine Learning course taught by Andrew Ng. The road map breaks down the topics for each lecture and includes timestamps for when each is discussed. CS229 gives you a good baseline for learning more advanced machine learning topics (guides on these are currently a work in progress).
The modeling road map will soon include practical examples for training your own models outside of CS229 and more comprehensive guides for advanced machine learning topics you may be interested in. Check out the Modeling Road Map here.
Developer Road Map
I've had many developers reach out to me to better understand machine learning because the recent advancements have made it plausible for anyone to build machine learning into their applications. My suggestion for developers is always to use the Machine Learning Engineer Road Map and skip the topics they're familiar with. While it goes deeper into the math than some developers think they need, I think even building with ML properly requires an understanding of how ML works.
However, I do understand that many developers learn by building, so I included a separate road map that runs through many of the resources needed for the MLE road map, but with a different approach. The approach is to find a problem you want to solve with an LLM and to try solving it with just ChatGPT, then using an LLM API, and finally using a model you've trained yourself.
This is a practical way to learn ML while building and also helps developers understand how a knowledge of machine learning is helpful for building ML products. Here's a link to the Developer Road Map.
Consumer Road Map
This is the shortest and simplest learning path, but the most important. Every single person will need to understand the role ML is playing in their life and how it affects them. Pretty soon every consumer will use AI daily to make their life easier. An understanding of how ML works will not only help you use AI, but it will also put you ahead.
The Consumer Road Map walks you through a basic understanding of machine learning principles and why machine learning will have such a large impact on consumers. It'll help you understand why you need to know ML and how you'll see it impact your life. You can find it here.
Updates to the consumer road map are coming soon to elaborate on the topics currently in it. This has been by far the most difficult road map to construct. While there are many great, free resources for the learning the science behind ML, there are much fewer for understanding how it impacts humans. Part of this is because it hasn't been as in-demand until now and part of this is because we don't fully know how ML will impact us.
Company Road Map
Even though I'm a machine learning engineer and I work for a company, this is the road map I have the least experience with. Many companies want to use ML, but don't know where to begin. Non-tech companies have many ways they can capitalize on ML to provide a better product to customers and save costs, but they have no knowledge to implement (or even identify) these ways.
My suggestion for any company in this position is to hire an AI consultant. While AI consultants can be expensive, a properly chosen AI consultant will more than pay off the cost that goes into them. I suggest reaching out to
if you (or your company) are interested in this. He’s an AI consultant who also helps other understand AI through his writing.More information can be found in the Company Road Map.
Coming Soon to Machine Learning for Everyone
I'll be keeping Machine Learning for Everyone updated over time with advancements in the field of AI. I'll also improve it to make it easier to use and more beneficial for everyone. Coming soon to Machine Learning for Everyone:
A glossary for ML terms. If you ever see a term you don't know, you can easily look it up in this glossary.
A learn-by-topic approach. You can choose a ML topic and find high-quality, free resources to walk you through it.
Practical machine learning examples anyone can run through. Notebooks and examples to help you learn ML hands-on.
An explanation of ML tools available. I've run into many ML tools recently I didn't even know existed but find very helpful. This will document all of them.
The best way to support Machine Learning for Everyone is simply by starring the repository on GitHub. If you have a GitHub account, this would be super helpful. You can also share it yourself.
The other ways to support it are following me on X and joining Society's Backend so I can share it with more people.
Thanks for sharing the roadmaps!