Classical AI: Search, Games, and Constraints

A build-it-yourself course: no jargon-memorizing, no quiz-grinding. In every session an
AI mentor walks you through a few dozen lines of pure-stdlib code and the algorithm runs
before your eyes — BFS flooding a maze like a ripple, A*'s expansion squeezed into a
narrow band by a heuristic, alpha-beta cutting half a million game nodes down to under 4%.
The course follows the topic skeleton of Berkeley CS188 — state-space search, heuristic
design, adversarial games, constraint satisfaction, all the way to Monte Carlo Tree
Search — with every script, maze, and game board rewritten as original hands-on teaching
for the online sandbox. By the end you'll see that AGV obstacle avoidance, game AIs that
read three moves ahead, and LLM planners (ToT/MCTS) all grow from the same search tree —
one you have planted yourself.