Before You Code

Before You Code: The Interview Questions Interviewers Expect You to Ask

Most candidates rush into writing code the moment they hear a problem. The best candidates don’t. They pause, clarify, and ask sharp questions that show they understand real-world engineering trade-offs. Here are the questions interviewers expect you to ask — and why they matter.

---

🔑 1. Input / Output Clarification

• Format: "Are inputs given as an array, linked list, or stream?"

• Constraints: "What are the input size limits (n, value ranges)?"

• Edge cases: "Can the input be empty? Can it contain duplicates/negative numbers/nulls?"

• Output: "Do you want all possible answers or just one valid answer?"

---

🔑 2. Functional Requirements

• Exact definition: "When you say 'sorted', do you mean ascending, descending, or any consistent order?"

• Ambiguity check: "When you say 'duplicate', do you mean exact same value, or same object reference?"

• Special cases: "What should I return if no solution exists?"

---

🔑 3. Performance Expectations

• Time complexity: "Is an O(n^2) solution acceptable, or do we need something more efficient?"

• Space complexity: "Can I use extra memory like a hash map, or should I optimize for constant space?"

• Realistic trade-offs: "Since input size can be 10^6, should I consider streaming or batch processing?"

---

🔑 4. Example-driven Validation

• Walk through examples: "For input [1,2,3,4], should the answer be ___?"

• Corner cases: "What about input with only one element, or very large values?"

---

🔑 5. Environment & Assumptions

• Language features: "Can I use built-in sorting, or do you want me to implement it?"

• Data structure availability: "Can I assume I have access to hash sets/dictionaries?"

• Mutability: "Should I modify the input in place or return a new structure?"

---

🔑 6. Problem Extensions (senior-level bonus)

Strong senior candidates often go one step further:

• "If input data arrives as a stream instead of a batch, how would the solution change?"

• "What if this had to run in a distributed system with multiple machines?"

• "Do we need to optimize for read-heavy vs. write-heavy workloads?"