AI Engineer Roadmap
Entire detailed AI Engineer roadmap
A friend of mine recently switched from Google to Meta for an AI Engineer role with a total compensation of over $500,000. And over the weekend, I was in a conversation with him to pick his brains on how he landed the job, how he prepared for it. And he said , he did it in three phases.
Learning about Data Science and Deep Learning :
He was good at coding but he didn’t know anything about data science. But when ChatGPT came out, he realized the AI boom and knew he had to join the bandwagon. But he needed a strategy. So, here’s what he did :
He used this github resource which was very helpful for him and was a one-stop shop for his Data Science Preparation.
In this resource, he started with this :
2 weeks : Machine Learning Courses for Free : Link
First, he started off by learning Machine Learning for free through the above link and learn supervised , unsupervised, regression models.
4 weeks : Machine Learning Projects : Link
Next, he wanted to implement the models hands on. So, he use the above link to prepare projects which he could not only gain experience but also learn to work with ML model hands on.
2 weeks : Andrew NG Notes Collection : Link
Then after 6 weeks, when he felt comfortable with ML, he focussed on Deep Learning. This is where he used Andrew NG notes collection which were super helpful.
4 weeks : 20 deep learning Projects with Python : Link . He said he implemented atleast 12 of these and knew them very well. He also customized them as per his learnings to truly own them.
4 weeks : Building Chatbots : Link
Next, he spent the remaining 4 weeks building chatbots using the above link to actually work with the AI models and train and improve them. This helped him understand how to build chatbots, how the AI models worked, how to tune and improve them. This information helped him get into the Google’s Gemini team.
At this time, he interviewed with Gemini’s team and was able to get into the core infrastructure team that actually built the framework for the Gemini’s AI models. This AI model was called Bard at the time.
Projects while at Google Gemini:
At Google’s Gemini, he worked on core projects to improve chatbot results as well as AI image generation and Notebook LM. While he couldn’t share exact details of the job due to its NDA status, he did share some of the things that helped him learn.
Deep Learning Neural Networks : Link
Deep Learning in Production : Link
Interview Prep :
Now, after 2 years of work, he was finally felt he had a strong base and experience to crack the AI Engineer roles at other companies and his motivation was to learn what other companies were doing with the AI race. His aim was to get into Open AI but he couldn’t clear the interviews. He got into Meta along with three other offers Big Tech. He joined Meta as it offered him the best compensation and a chance to work on core AI models like Llama.
Here was his roadmap
There were 4 key interviews he prepared for :
Coding
AI model and Deep Learning
System Design
Behavioral
Coding :
He was already a good programmer. So, he just had to revise his leetcode skills. For that, he used Blind 75 as a start but he didn’t stop at that.
He wanted to ensure he cleared his coding interview even for hard questions so he also solved. So, he did top most asked questions for the below companies :
Open AI leetcode : Link. This is where he really wanted to work.
Meta Top Interview Questions : Link. He only solved top 100 most liked or most upvoted questions.
AI Model and Deep Learning :
For this, he used the below resources to prepare for the theoretical questions that could be asked :
Challenges and Application of LLM
AI Engineer Interview Questions
System Design Question
He already knew basic System Design for which he had used 2 resources :
Alex Xu - System Design Interview : He bought this book. Pdf
Along with this , he studied big Data Concepts. For that, he used this resource for interview questions.
Behavioral :
Believe it or not, this aspect is where he actually struggled a lot. I actually helped him with this. And he loved how I had used the Amazon LPs to prep for it. All the questions are covered in this resource.
WEEK BY WEEK ROADMAP :
