Maximize Unique Letters

Given a list of lowercase words, choose any subset of words such that: - no selected word contains a repeated character internally, and - no character appears in more than one selected word across the subset. Return the maximum possible number of distinct characters covered by the selected subset. Follow-up stages: 1. Solve the exact problem for small inputs such as `n <= 12`. 2. Scale the solution to much larger inputs, potentially up to around `10^4` words. You should preprocess each word into a 26-bit mask, discard words that are internally invalid, and describe or implement pruning and a more scalable state-compression strategy.