High Performance Python
Practical guide to profiling, optimising, and scaling Python code — covers NumPy, Cython, asyncio, multiprocessing, and memory management in depth
Visit High Performance Python →Goes past “use NumPy” into actually profiling code to find real bottlenecks first, then covers the specific tools for fixing them — Cython for compiling hot paths, multiprocessing and asyncio for concurrency, and memory-layout considerations that matter once a program is processing enough data for memory to become the constraint.
The profiling-first framing is the book’s core discipline: optimizing before measuring is how developers waste time speeding up code that was never the actual bottleneck — this book is built around avoiding exactly that mistake.