Performed Evolution Strategies to solve Blackbox optimization problems. Implemented various variance-reduction techniques including Vanilla sampling, Antithetic Variate, Gaussian Orthogonal Matrices, and Random Hadamard Matrices. Also implemented regression-based gradient estimators including LP-decoding, Lasso, and Ridge. Tested the implemented Evolution Strategies algorithm on reinforcement learning tasks from OpenAI Gym. Analysis shows ES algorithm with regularized regression and antithetic variate can estimate the gradients of unknown functions well and lead to high rewards.