Trading is rigorous, mathematics is rigorous, coding is rigorous, and quantification is the most rigorous of all; this is not a child's play area.

Have multiple sets of cross-comparisons been long-term tested? Is the testing period 180 days or 360 days? Has it experienced extreme market conditions, black and white swans? For accounts started at different times, do the returns vary significantly? If all accounts enter and exit at the same point, can the liquidity support it (slippage probability)?
If you're just playing for fun and entertainment, anything goes. But if the plan is commercialized, these questions should have been considered early on; otherwise, it will definitely be a nightmare later.
Save some of the time spent bragging every day, read a few more professional books, don’t always think AI is all-knowing and omnipotent. When designing the framework, think more comprehensively. Don’t always hold on to the idea that, “Oh, I wrote this (or AI wrote this), and I can make trades, I’m so awesome.” Or that the mountain of crap downloaded from GitHub runs smoothly and can make trades, and I’m so incredible.
This is not quant research; this is stupidity, it’s showing off foolishness.
However, everything is a two-way pursuit; just like a telemarketing call with a Fujian accent, if you believe that, then it can only be true love.