Day 16 of 100 Days Agentic Engineer Challenge: Progressive Learning
The most common scenario regarding doing some course is problem of time to take the course, to practice it. Many people start with few first lessons, there is a motivation at the beginning, then again they leave it for later. The nice idea would be to start with an overview of whole material on the first day and then continue with lessons but in basic version, then repeat in intermediate version and at the end master the topic in advanced level. This would give us the whole course experience from day one and in each level we would also get the whole content but at different levels. So it wouldn’t be a problem to take a break after one week, because after the first level we get the basic understanding. I’ve built something to simulate this process, but first let’s review my daily tasks routine.
Daily tasks routine
1.Physical activity — It’s my priority, started slowly, from 0 to 25 yesterday, and it’s the 15th day of my journey, I reached 100, not days, but push-ups ;)
2.Seven hours of sleep — Every night I think I should go to sleep earlier and it is slowly starting to work.
3.AI agents — I learn about agents every day, watch YT, need to build one, a simple one, but one that will work autonomously 24h.
4.AI assistant course: Yesterday, instead of doing one lesson at a time, I downloaded the full source code and ran the app, I just changed the AI provider from OpenAI to Azure OpenAI. I actually applied the progressive learning method and tried to get everything in the shortest time possible to see the complete course results before going into individual lessons. Then I started reviewing the code associated with each lesson to get a basic understanding of the first level. I have to admit that this way of learning works great.
5.PAIC: in queue
6.Data Science: in queue
If you want to know what all these tasks are about, read the introduction to the 100 Days Agentic Engineer Challenge.
Personal Learning AI Agent
I’m not sure we can make an AI learning agent, because agents usually do things for us, and in terms of studying, they can learn for us for exams, but they could be helpful in monitoring our learning activity to help us acquire knowledge in the best way possible. Everyone is different and learns in different ways and individual optimization could be interesting. Another thing is to remind all the time about upcoming exam, importance of finding time and coming up with ideas to just look up few topics. The biggest problem of most people is to start learning too late, then there is too less time and acquire enough knowledge what starts to be very stressful. The solution would be to convince the person to start learning early enough, so he can learn every day but short amounts of time during longest period.
Progressive Learning Method
I made a simple PoC app. You give a topic like machine learning or some smaller topic like photosynthesis and the app generates the knowledge graph and 3 levels of difficulty of the same content with quizzes. The idea is, as I mentioned before, to go through the whole content at a first time looking at content like in the example a knowledge graph. We see very topic, subtopic on the graph and we know how much material we need to acquire, then we start to study level one and again we have the whole scope in the basic version, then again the intermediate version and once again the advanced one, so four times we learn the same from different point of view.
I think this is the best way to learn anything, with continuous building of deeper understanding. The most important idea is to go through the whole content as fast as possible, even if we are not able to get much the first time. So if we’re making a coding video course, let’s watch each lesson in speed mode without practicing it, and the second time look for details and maybe repeat it the third time. I was already able to use this method for one course as I describe in my daily tasks routine part, and will use it again for the next ones. Happy learning :)
