Adapt to interviewer coding preferences

High‑level approach - Calibrate in the first minute: confirm problem type, constraints, and success criteria. - Narrate your plan: restate, propose, justify, implement, test, and iterate. - Deliver a minimal, correct solution first, then optimize. A) Algorithmic problem 1) Preparation - Practice core patterns: two pointers, sliding window, hash map counting, heap/priority queue, BFS/DFS on graphs, union‑find, binary search, DP. - Drill time/space trade‑offs and Big‑O for common data structures (array, hash map, set, heap, queue, stack, trie). - Practice writing clean code without IDE help, and speaking in a structured way. 2) Clarify requirements (example prompts) - Input: exact format and size bounds (n, range of values). Are negatives, duplicates, or empty inputs possible? - Output: any valid solution or a specific ordering? Deterministic tie‑breaking? - Constraints: time/space targets; can we use extra memory? Streaming input or all in memory? - Edge semantics: off‑by‑one cases; sorted vs unsorted; is the graph directed/weighted? Can there be cycles? 3) Choose data structures: concrete example Example task: Return the k most frequent elements from an integer array. - Reasoning: we need counts and a way to keep the top k efficiently. - Data structures: hash map for frequencies; min‑heap of size k to retain top k by frequency. - Algorithm outline - Build freq map in O(n). - Push (freq, value) into a min‑heap. If heap size > k, pop the smallest. - Extract heap contents for the answer. - Complexity: time O(n log k), space O(n). - Small example - Input: [1, 1, 1, 2, 2, 3], k=2. - Freqs: {1:3, 2:2, 3:1}. - Heap after processing: [(2,2), (3,1)] -> output [1, 2]. - Alternatives - Quickselect on frequencies for average O(n) time and O(n) space (more complex to implement under time pressure). 4) Validate edge cases - n=0 → return empty list. - k=0 → empty list; k > unique count → return all uniques. - All items same frequency → any k items acceptable unless order required. - Large n → confirm O(n log k) meets constraints; consider quickselect if n is huge and k large. - Memory constraints → discuss streaming approach only if required. 5) Communication script snippets - "Let me restate the problem to confirm understanding..." - "Given n can be up to 1e5, O(n log k) with a min‑heap seems safe; alternative is quickselect for average O(n)." - "I'll test empty input, k=0, and ties in frequencies before finalizing." B) Hand‑written front‑end code 1) Preparation - Refresh DOM APIs and patterns: document.querySelector, addEventListener, classList, event delegation, focus management, form controls. - Browser APIs: Fetch + AbortController, setTimeout/requestAnimationFrame for debounce/throttle, IntersectionObserver (if needed), localStorage/sessionStorage. - Accessibility basics: roles, labels, keyboard navigation, focus order. - Performance: avoid unnecessary reflows; batch DOM updates; handle async races. 2) Clarify requirements (example prompts) - Environment: vanilla JS only? Any library allowed? HTML/CSS available or JS‑only? - Functionality: must handle keyboard, mouse, and screen readers? Mobile support? - Data: API shape, response time, pagination, errors, rate limits, debounce requirements. - Performance and UX: max items to render, loading states, empty states, retry behavior. 3) Choose browser APIs: concrete example Example task: Build a debounced search box that fetches suggestions and allows keyboard navigation. - Minimal structure (describe): an input with id='q' and a container ul with id='results'. - APIs and patterns - Input handling: addEventListener('input', ...). - Debounce: setTimeout/clearTimeout (e.g., 300 ms). - Networking: fetch(url) with AbortController to cancel in‑flight requests on new input. - Rendering: createElement/li, textContent, appendChild; clear via innerHTML=''. - Accessibility: role='combobox' on the input container, aria‑controls='results', listbox options with role='option', aria‑selected; manage aria‑expanded. - Keyboard: handle 'ArrowDown', 'ArrowUp', 'Enter', 'Escape' on the input; maintain an activeIndex; ensure focus management. - Event delegation: click on the ul to detect which suggestion was clicked. - Data flow - On 'input': if value is empty, clear results and set aria‑expanded='false'. Otherwise, set a debounce timer. - When debounce fires: abort previous fetch via AbortController, fetch suggestions, handle errors. - On success: render up to N items, update ARIA attributes, reset activeIndex. - On 'keydown': update active selection with arrow keys; on 'Enter', commit selection. - Guarding against races - Use AbortController or a monotonically increasing requestId to ignore stale responses. 4) Validate edge cases - Empty input, whitespace‑only, or repeated same query. - Slow network or errors: show 'loading' and 'no results' states; retry guidance. - Stale response arrives after a newer one: ensure it is ignored. - Zero results: render an empty state, maintain aria‑expanded='false'. - Long or international input: composition events; ensure string normalization if needed. - Keyboard navigation: wrap at ends; Enter should select; Escape should close. - Memory and cleanup: remove or reuse listeners if component is disposed; avoid leaking AbortControllers. - Security: escape textContent (avoid innerHTML) to prevent XSS. 5) Communication script snippets - "To confirm, vanilla JS only, and we need keyboard accessibility with ARIA roles?" - "I'll debounce input to 300 ms and cancel prior requests with AbortController to prevent stale updates." - "Let me test: empty input, zero results, network error, fast typing that triggers cancellation, and keyboard navigation." Bridging both scenarios: how I switch gears - First minute: confirm scope and constraints specific to the interviewer’s focus. - State a plan aligned to that focus (data‑structure‑first vs. DOM/API‑first) and ask for buy‑in. - Build the simplest correct version, then iterate (optimize algorithm vs. enhance UX/performance/accessibility). - Keep a running test list and check items off audibly to demonstrate correctness and ownership.