LeetCode AI Related Problems

[pertrai1]

Gather a list of problems in LeetCode that can be considered AI related. Update this issue with the list that has been collected.

AI & LLM-Oriented LeetCode Problems

Curated LeetCode problems that reinforce algorithms and data structures used in modern generative AI/LLM systems: tokenization, sequence alignment, attention-style windows, retrieval, and sparse linear algebra.

Tokenization & language primitives

• 0208 - Implement Trie (Prefix Tree) — prefix vocabulary construction for token lookup/autocomplete. - My Solution
• 0212 - Word Search II — trie-guided backtracking similar to constrained decoding. - My Solution
• 0139 - Word Break — DP segmentation akin to validating token boundaries (BPE-like). My Solution

Sequence alignment & generation quality

• 0072 - Edit Distance — Levenshtein alignment for evaluation and decoding heuristics. My Solution
• 0115 - Distinct Subsequences — sequence likelihood/counting via DP. My Solution
• 1143 - Longest Common Subsequence — overlap/ROUGE-style similarity between sequences. My Solution

Attention windows & context management

• 0076 - Minimum Window Substring — minimal span covering required tokens (attention window intuition).
• 0340 - Longest Substring with At Most K Distinct Characters — fixed-size vocabulary within a context window. My Solution

Retrieval & similarity search

• 0692 - Top K Frequent Words — vocabulary curation and frequency-based ranking. My Solution
• 0658 - Find K Closest Elements — 1D k-NN building block for ANN-style retrieval. My Solution
• 0973 - K Closest Points to Origin — vector similarity search (Euclidean/cosine analog).

Sparse linear algebra for transformer kernels

• 0311 - Sparse Matrix Multiplication — efficient matmul on sparse activations/weights.
• 1570 - Dot Product of Two Sparse Vectors — embedding similarity with sparse representations.

Graph reasoning for semantic transitions

• 0127 - Word Ladder — shortest-path search between tokens/words using BFS.