Anthropic coding interview
Round 5 of 8 · 18 questions reported
- Length
- 55 min
- Format
- Live coding on a realistic, multi-part problem, a different one from the screen
- They grade
- Complete, runnable code; senior candidates are expected to think of the test cases themselves
- AI use
- No AI
- Heads-up
- Usually one or two coding rounds across the onsite. Some 2026 loops add a separate round where you write and review code with Claude Code in a hosted editor; there AI use is expected.
Questions Anthropic asks in this round 18
- Deduplicate files by size then content hashSep 2026 · very often
- Image transform pipeline with a wall-clock budgetSep 2026 · very often
- Reconstruct call stacks from enter and exit tracesSep 2026 · often
- Concurrent same-host web crawlerSep 2026 · often
- Multi-step task with an LLM tool-use loopSep 2026 · often
- Fix a function-cache key function, then make it durableJun 2026 · often
- Distributed mode and median across worker shardsSep 2026 · sometimes
- Greedy longest-match tokenizerJun 2026 · sometimes
- Weighted data batcher with deterministic resumeMay 2026 · sometimes
- Repair a looping bootloader programJul 2026 · rarely
- Kernel-optimization take-home scored by speedupDec 2025 · rarely
- Banking system with transfers, cashback and account merges
- Cloud storage with user quotas and backups
- Group prompts into templates under concurrent ingestion
- In-memory note-taking system with keyword search
- Minimum workers with first-free job assignment
- Task manager with user quotas and expiring assignments
- Thread-safe task queue on a linked list
About the Anthropic coding round
What does Anthropic ask in the coding round?
The most asked: Deduplicate files by size then content hash, Image transform pipeline with a wall-clock budget and Reconstruct call stacks from enter and exit traces. 18 questions reported in all.
How long is the Anthropic coding round?
About 55 minutes.
What is the Anthropic coding round like?
Live coding on a realistic, multi-part problem, a different one from the screen. They grade: Complete, runnable code; senior candidates are expected to think of the test cases themselves.
Can you use AI in the Anthropic coding round?
No AI.