HN user
Well, they're usually revising/improving their models while the robots print money today to ensure that the robots can print more money tomorrow. The frequency and depth of the human input varies based on the kind of strategies employed (macro/statistical arbitrage etc). Moreover, successful models take competitors' potential models into account which (like everything else) is a moving target.
> At the end of the day, I'm trying to answer the question: can you make money in the stock market with a low-touch (eventually no-touch) approach with the help of advanced software and analytics
The answer to this has been a resounding "hell yes" for decades across IB/HFT/AlgoTrading shops. The only thing this exercise achieves is whether "you" personally can make money with a low-touch approach using software.
This. From OP's blurb it's pretty obvious he's very new to this and most likely hasn't traded much stocks even using the "simplistic human investing" approach. If you're planning on using this exercise to help "mature" your understanding of how you should start tracking/investing stocks then this could be a rewarding effort. If, however, you're planning on jumping straight in to build this to "actually" trade stocks I'm afraid you'll be in a world of hurt before you even know it. If you're serious about the journey, then the choice of programming language or DB is a long long way down the road; start from the basics and work your way up from there.