The slowdown of Moore’s law, then the rise of multicore and heterogeneous specialized computing architectures and the new generation of programming languages or frameworks create an opportunity to bring perfect harmony into the AI/ML computing at edge, cloud and data centers.

MulticoreWare has a team of compiler engineers with experience and expertise in areas like Heterogeneous Programming Languages and Runtimes (CUDA, OpenCL, Renderscript etc.), domain specific languages (Caffe, TensorFlow, PyTorch, Halide etc.), Neural Network optimization engines (TensorFlow XLA, TVM, GLOW etc.). MulticoreWare provides licensable IP (MxPA), engineering services in developer toolchains (debugger, profiler, OS/standard libraries) amongst many other related engineering consultancy services.