Today, computing is almost exclusively done via the Digital Computation paradigm (DC) based on Turing machine theoretical model. Also, Quantum Computation (QC) paradigm has recently started to gain momentum. A radically different ''Random Pulse Computing'' (RPC) paradigm, whose principle was initially proposed in the 60's, makes use of counting and combining electrical pulses that appear randomly in time, similar to nerve pulses and inspired by information processing in neurons of living beings. Revisited in the past decade, RPC has attracted high attention because of it economic use of resources, fault tolerant computing, low power operation and yet a promise to be able to efficiently solve artificial intelligence problems that are hard for both DC and QC, such as: image processing, deep learning and problem-solving that mimic human brain operation. We propose a research towards a general theory of synthesis of circuits for a universal RPC computer based on the notion of entropy. Based on that, we will demonstrate a programmable Random Pulse Computer, and a use-case in image processing, built on a chip using standard Si CMOS process and quantum randomness of single-photon detection. We will build upon our recently published works on Entropy Budget Criterion (EBC) and random flip-flop, which together present a powerful tool for synthesis of novel RPC circuits. We are excited by the fact that bio-inspired sensing occurs naturally in the technology of single-photon counting camera, a successful research topic at AquaLab at EPFL, where indeed each light sensor (pixel) encodes information in a random pulse train of quantum origin. Information in that form can be processed directly by RPC circuits, a major research topic at RBI. In mutual collaboration, we have recently built a World's first random flip-flop on a chip carrying 2800 T-type RFFs. This new technology is a crucial asset for reaching the goals of this project.