How did my AIAP selection round go?

Last weekend, I had the opportunity to participate in the selection round for the Artificial Intelligence Apprenticeship Program. This is one of the most challenging try-outs I have ever attempted, because to qualify for this round, I must successfully write a program that employs three machine learning models to process a dataset. 

For this dataset, I need to predict the activity levels of seniors living in a habitat equipped with various sensors, including those measuring carbon dioxide, carbon monoxide, and humidity levels. I completed the programme within the deadline, but my accuracy was about 59%, which is not particularly good. Nevertheless, my submission made it to the final round.

In this final round, I had to first present my solution and then undergo a technical interview. After that, I would be grouped with other candidates to solve an extension of the earlier problem set that I had already solved. The final exercise was really exhilarating, as two hours is too short to run a new dataset through even one machine learning code; making matters worse, we still needed to shape the output in a form that is acceptable to the invigilator in the form of a Python program. This really reminds me of a concept known as the "Kobayashi Maru" from Star Trek.

Furthermore, I suspect that I was grouped with the "uncles" team, basically the oldest folks in this batch, but ultimately this was a good thing, because everyone was calm and mellow. I was tasked with cracking the data cleaning part. At the same time, a buddy prepared his Python program to accept my output, and we barely squeezed out a solution at 88% accuracy, 15 minutes before time was up. 

During debriefing, the investigators had to ask more probing questions about what we understood about the problem and how we thought the ideal approach should be taken to crack it. The rumour was that 50% will be culled in this round, so the final batch that qualifies for deep skilling in AI will be small. 

In an ideal world, this is what skills training should be. The starting point is an asynchronously delivered lesson, and to qualify for deep skilling, you need to produce a functional product to advance to the next level after attending a short course. After which, you still need to get together with others to see whether you can coordinate with a team to get real results. If you want to prioritise skills over qualifications, then to qualify for skills training, there must be a high rejection rate. 

My results, which will be announced by the end of October, are ultimately in the hands of the judgment panel, and I've done my best. If my best is not good enough, so be it, as it's been a challenging and rewarding contest.

For this upcoming week, I will be extending my discussion on covered call strategies. The last video on covered calls did not receive a decent amount of views, but it had great interaction, so I conducted further backtesting to address some questions from my viewers.

On Thursday, I take my biggest risk ever as I deviate from my ordinary discussions to talk about Incel life and the Manosphere. I discuss how to raise boys without them turning into incels and the possibility of financial education that can direct men's energies into more positive pursuits.

As I discuss thought leadership in the Manosphere, such as Andre Tate, YouTube has flagged some issues with my content and reduced monetisation opportunities, but I have appealed that decision. Therefore, I may not be able to discuss this in the future.