An ESP32 seems a better fit to me, and would also be cheaper and more power efficient.
Still, I would politely suggest that you're underestimating the part where we needed low level access to a camera module that came with the RPi.
Even if the ESP32 had been out, we didn't have anyone handy to spend six months reverse engineering CMOS registers.
Everything is super easy a) in hindsight and b) when you're commenting on Hacker News and not actually doing it.
(And I would politely ask that you don't make assumptions about what I might or might not have actually accomplished in similar spheres, you have zero idea.)
Happy to answer any further questions you might have to the degree that I'm able.