These aren't reference docs. Each topic starts with something you can poke at — an interactive demo that builds the intuition before any code appears. Pick one and give it five minutes.
Walk both ends toward the middle and see what falls out.
The pattern that makes subarray problems click — fixed-size or shrinkable.
Trade space for speed. O(1) lookup changes what's possible.
Last in first out, first in first out. Most problems reduce to picking the right one.
Halve the search space every step — not just for sorted arrays.
Nodes and edges. Most tree and graph problems boil down to traversal order.
Solve it by solving a smaller version. Undo what doesn't work.
Overlapping subproblems, optimal substructure — the real intuition behind DP.
Wait until they stop, or fire at most once per interval. Two sentences, two patterns.
Same input, same output? Cache it. useMemo and useCallback are just the React version.
One handler on the parent, not fifty on the children. Events bubble — use it.
The event loop, microtasks, and why your setTimeout fires after your Promise.
Implement the thing without using the thing. The interview question that tests understanding.
SSE, streaming, state machines, and the UI components that make agent features work.