Learn about an innovative AI model optimization toolchain in this 20-minute conference talk from tinyML Asia 2023. Discover how PyNetsPresso, a Python-based interface for NetsPresso, streamlines AI model development by automating training, lightweighting, model conversion, and benchmarking processes. Explore the capabilities of LaunchX, a GUI tool that enables visualization of runtime and hardware configuration impacts on model performance. Master hardware-aware optimization techniques including pruning, filter decomposition, and quantization while understanding how to minimize performance degradation under hardware constraints. Gain practical insights into developing optimized AI models for specific hardware configurations through an integrated toolchain that simplifies the complex iteration process and enables efficient benchmarking across various hardware setups.