A graduate-level reference that unites rigorous mathematics with hands-on computation. Twenty-four tightly written chapters carry the reader from floating-point arithmetic to large-scale parallel solvers, always pairing theorems and proofs with annotated Python code. Why this book? • Comprehensive coverage of LU and Cholesky factorization, QR decomposition, and Singular Value Decomposition (SVD) – the staples of every scientific computing and machine learning stack. • Complete treatments of iterative methods such as Conjugate Gradient, GMRES, and Lanczos-based eigenvalue algorithms, including advanced preconditioning strategies. • Up-to-date material on randomized linear algebra, low-rank approximation, and sketching – indispensable for modern data science pipelines. • Detailed chapters on GPU acceleration, communication-avoiding algorithms, and distributed memory implementations , giving readers a clear path from theory to high-performance code. • In-depth discussion of condition numbers, backward error analysis, and stability , providing the mathematical guarantees demanded in engineering and quantitative finance. • Every chapter closes with ready-to-run Python notebooks that reproduce all numerical examples and visualizations. Key contents Vector norms, spectral radius, and condition numbers - IEEE floating-point and roundoff analysis - Backward stability of Gaussian elimination - Blocked and communication-optimal LU, QR, and Cholesky - Least-squares, Tikhonov regularization, and linear regression - Power, inverse, and Rayleigh quotient iterations for eigenvalues - Bidiagonal SVD algorithms and sensitivity results - Krylov subspace methods – CG, MINRES, GMRES, BiCGStab - Preconditioning, algebraic multigrid, and spectral transformations - Matrix functions – exponential, logarithm, and fractional powers - Low-rank approximation for data compression and machine learning - Randomized matrix multiplication, CUR, and RSVD