
How to get into AI?
Want to get into AI? You'd like to ultimately get certified? Here are some resources to help you with that.
Jean-Francois Landreau
Amazon Employee
Published Oct 23, 2024
Last Modified Oct 25, 2024
Thanks to ChatGPT, AI, ML and Generative AI skilled profiles are in high demand. However, it may still be difficult to get started for a person that has not studied these topics at the university. It's my case. Here is how I proceeded.
You need to establish some foundations. A few years ago, before the generative AI craze, I took a course about machine learning from Andrew Ng. He is certainly one of the most prolific educator on the topic. This course is going to give you the vocabulary and a basic understanding of what is Machine Learning. Still you'll have enough knowledge to start to apply Machine Learning on simple scenarios. And you'll probably know more than most people talking about Machine Learning on the Internet.
Once you have the foundations, you can start to add the generative AI layer on top. In the end, it's a branch of ML with specific algorithms and specific ways to train the models. Also the usage of these models are different but you have certainly played with them already and you know understand what to expect from an end-user perspective. Here again, Andrew can help with this course: https://www.coursera.org/learn/generative-ai-with-llms. You'll gain the vocabulary and a basic understanding about what is behind these generative AI chatbots and how they are produced. And gain, you'll probably know more than most people talking about generative AI on the Internet.
Generative AI is the most active part of the Machine Learning universe and it's moving fast. So it's good to follow a few persons to get update.
Here are a few other good profiles to follow on Linkedin:
- Philipp Schmid gives updates on the new models and new studies. It's good to have an overview of the generative AI ecosystem without the bias from the big players.
- Eduardo Ordax publishes fun memes and articles to learn at high level what's happening without getting too deep into the details.
- David Sauerwein gives updates on research papers expanding the horizon above the surface where most people crawl.
- Mike Chambers runs some of the trainings on generative AI with Andrew Ng and posts regularly fun videos with updates and tips & tricks.
AWS certifications exams ask questions related to AWS Services but the knowledge required to answer most questions can be applied in other contexts. So it's always a good validation whether you know the domain well enough or if there are still some areas where you need to learn further.
For the AI Practitioner certification, many question are related to generative AI, so I recommend to go though the Building with Amazon Bedrock workshop.
For the ML Engineer Associate certification, more questions are about Amazon Sagemaker, so I recommend to go through the Sagemaker Immersion Day.
In both cases, the certification guides points to documentations that are good to read before the exams:
Another thing to absolutely do before the exam is to run the test exams on Skills Builder:
The number of resources on AI/ML and generative AI is huge nowadays. So you can certainly find other courses and go deeper in many directions.
One topic I find interesting is to understand better the algorithm behind the generative AI models and there is nothing better than looking in the code to do that. Here is an article that does that with the model that started the GPT(Generative Pre-trained Transformer) trend: BERT.
I've been working in IT for many years and I could re-use my experience when going through this learning exercise. Also I have collected many years of experience on AWS which is useful particularly for the exams that cover not only the AI/ML and generative AI services but require some general AWS knowledge as well. Finally, in my daily job, I can work on projects related to AI/ML and generative AI. There is nothing better than learning on the job. But get started, take your time and find ways to practice. And if you provision a large GPU cluster, it may be costly. So set a budget and an alert on the AWS account where you train.
Any opinions in this post are those of the individual author and may not reflect the opinions of AWS.