Each track is a set of chapters, and each chapter is a pattern plus the problems that drill it. A track unlocks when every prerequisite is at 80% or better.
Indexing, in-place mutation, prefix sums, and the string operations that show up everywhere else.
Converging, parallel, and fast-slow pointers. The first real pattern, and the one everything else borrows from.
Trading space for time: frequency maps, seen-sets, grouping, and when hashing quietly degrades.
Fixed and variable windows, and the invariant that tells you when to shrink.
Reading a call stack, writing the base case first, and pruning search trees before they explode.
Binary trees, BSTs, and the four traversals — plus the recursive shape that solves most tree problems in six lines.
Sift up, sift down, top-k, and the streaming problems that only a heap can answer.
Representation, BFS and DFS, cycle detection, topological order, and connected components.
Dijkstra, Bellman–Ford, and Floyd–Warshall — and choosing between them under real constraints.
Memoisation to tabulation, state design, and the eight recurrences that cover most interviews.
Exchange arguments, interval scheduling, and proving a greedy choice is actually safe.
Tries, union-find, segment trees, and Fenwick trees — the structures that turn O(n) queries into O(log n).