Bayesian Optimization: A Pure-Python Sequential Model-Based Optimization Library

We present a lightweight Bayesian optimization library built on NumPy and SciPy. The library maintains a Gaussian Process surrogate model and uses the Upper Confidence Bound (UCB) or Expected Improvement (EI) acquisition function to sequentially probe a black-box objective. On the standard benchmark f(x,y) = -x^2 - (y-1)^2 + 1, the optimizer achieves target value > 0.9 after 5 iterations starting from 5 random explorations, demonstrating rapid convergence with no gradient information required.