Differentiable Ray Tracer in Julia

Differentiable Programming is a programming paradigm in which we can differentiate through the program itself. It allows us to exploit the knowledge already embedded in a problem and use existing deep learning techniques to it. This project aims to develop a Ray Tracer in Julia and interface it with Flux/Zygote for Automatic Differentiation support. Also we shall demonstrate the use of this Ray Tracer in downstream Reinforcement Learning Tasks.