30% of all my algorithms are trained based on various deep learning models. But it's certainly not chatgpt. If you understood what data it was trained on you would understand. What you are asking for is for tools and opinions without having done the slightest homework on your end.
How to get started? Like with everything in life: the boring, tiring, time consuming grunt work to familiarize yourself with the basics of the trade. In this case linear algebra, data science, basic ML, then deep learning models, and concepts of data engineering. 1 to 2 years to get up and running if you put a lot of effort into it and allocate a sizable chunk of free time. Are you willing to do that? Then I can guide you to suitable material. Any off the shelf tool that you can use after 15 minutes of trial and error has the exact value added of ZERO.