eBay·Software Engineer·Technical Phone Screen·Intermediate

Intermediate

Apr 2026

Summary

eBay SWE interview that focused on extending an in-memory file system across multiple layers of complexity. The final piece was a capacity eviction problem that required careful thought about ordering and data structures.

Questions Asked (1)

You're given an in-memory file system that already has basic file operations and per-user storage tracking. Implement an updateCapacity(userId, newCapacity) function that, if the new capacity is too small for the user's current files, evicts files from largest to smallest (alphabetical name as tiebreaker) until the total fits, and returns the count of evicted files. Return None if the user doesn't exist, 0 if no eviction is needed.

Algorithms & Data StructuresSystem DesignTechnical Trade-offs

Author's notes

The base file system stuff was fine, I'd seen similar problems before.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Clarify the data structures and requirements, then design an efficient eviction strategy using a max-heap or sorted list. Discuss trade-offs between time and space complexity, and handle edge cases like non-existent users and exact capacity matches.

Pro tip: Mention that you would keep files sorted by size and name to avoid re-sorting on each update, and use a heap for O(log n) evictions. Also, discuss how to handle concurrent updates or persistence if the system scales.

1. Clarify Requirements and Assumptions

Ask about the data structures available, whether file sizes can change, and if the eviction order is strictly largest-to-smallest with alphabetical tiebreaker. Confirm return values for edge cases.

2. Design Data Structures

Propose maintaining a max-heap keyed by (size, name) for each user, or a sorted list if updates are infrequent. Ensure per-user storage tracking is updated efficiently.

3. Outline Algorithm

Check if user exists; if not, return None. Compute total size; if <= newCapacity, return 0. Otherwise, repeatedly evict the largest file (tiebreak by name) until total fits, counting evictions.

4. Analyze Complexity and Trade-offs

Discuss time complexity: O(k log n) for k evictions with heap, or O(n log n) if sorting each time. Space complexity: O(n) for heap. Mention alternative approaches like balanced BST.

5. Handle Edge Cases and Extensions

Cover cases: newCapacity negative, zero, or larger than current total; user with no files; concurrent updates. Suggest optimizations like lazy deletion or batch updates.

Key Points to Mention

Use a max-heap or priority queue for efficient eviction of largest files.
Tiebreaker: when sizes equal, evict files in reverse alphabetical order (or as specified).
Update per-user storage tracking after each eviction to maintain consistency.
Return None for non-existent user, 0 if no eviction needed, else count of evicted files.
Consider time/space trade-offs: heap vs sorted list vs balanced tree.
Discuss potential concurrency issues and how to handle them (e.g., locking).

AI-generated suggestions, not part of the candidate's original notes. May be inaccurate — verify before relying on them.