Summary
Joe Jheng is a Senior AI Compiler Engineer based in New Taipei, Taiwan, with two decades of experience bridging real-time graphics, game engineering, and AI systems. He designs cross-layer compiler technology—from ONNX front-end conversions through MLIR-based middle-end fusion to VLIW back-end toolchains—enabling HW/SW co-design and faster silicon bring-up. Previously he led cloud gaming product strategy and technical delivery at Ubitus, shipped thousands of games across consoles and streaming platforms, and founded a mobile game studio that amassed over a million downloads. Joe combines low-level systems craftsmanship (assembler, scheduler, instruction generators) with developer-focused ecosystem work, having exposed runtimes via OpenAI-style APIs and integrated Hugging Face and llama.cpp. His background in game frameworks and performance-tuned engines informs pragmatic compiler optimizations that reduce operator counts and accelerate inference. Known for turning research-grade ideas into production toolchains, he excels at making specialized hardware accessible to mainstream AI workloads.
9 years of coding experience
19 years of employment as a software developer
Bachelor of Computer Science, Computer Science, Bachelor of Computer Science, Computer Science at National Tsing Hua University
Chinese, English