It just ended at the end of September. LinkedIn SDE Virtual Onsite, the overall experience was a bit smoother than I expected. The main reason is that I did the mock in advance and practiced the rhythm and logic of the questions, so I didn't seem to be flustered during the interview. The first round is a typical combination of coding + project deep dive, and I'll talk about it in detail below.
Opening & Introducing Yourself
The interviewer was a white male with a friendly demeanor. I was asked to do a self-introduction first at the beginning and I followed the usual order:
- Schools & Programs
- field of research
- Internship/Project Experience
We suggest that you don't just memorize a paragraph, but pick the key points. For example, when I talk about my project experience, I don't just say "I did a machine learning project", but specifically point out the application of the project and my main work. This allows the interviewer to find the point they want to pursue more quickly.
Sure enough, the interviewer then asked in-depth questions about my project:
- What are the main elements of the project?
- What specific techniques were used?
- What was the biggest challenge encountered? What did you do to solve it?
- What was learned from the program?
This part lasted about 10 minutes. The overall feeling was that it wasn't difficult, but rather that I wanted to see if you really understood the program, and if you could express your thoughts clearly.
Coding Question 1: Locker Switch Status
The first question is a classic one, and there are similar ones on Leetcode:
Problem 1: N lockers toggle problem
There are N lockers, all closed initially. On the first pass, you toggle every locker. On the second pass, you toggle every 2nd locker. On the third pass, every 3rd locker, and so on, until the Nth pass where you only toggle the last locker. On the third pass, every 3rd locker, and so on, until the Nth pass where you only toggle the last locker. Which lockers are open in the end?
The key point is the translation of the idea. The number of times each locker will be toggled depends on the number of factors it has. If a number has an even number of factors, it will eventually close; if it has an odd number of factors, it will stay open. Onlyexact squareIt's only an odd number of factors, so the last open locker is where all the squares are.
Once the idea is clear, writing the code is relatively straightforward. When I implemented it, I wrote a function to determine which numbers are perfect squares and then return the result.
In terms of time complexity, the final code can do O(sqrt(N)), because it is only necessary to iterate over the square numbers. The interviewer asked about the derivation of complexity and I explained it briefly.
Coding Question 2: Nested Integer Weighted Sum
The second question is also a common Leetcode question:
Problem 2: Nested List Weighted Sum
Given a nested list of integers, return the sum of all integers in the list weighted by their depth. Each element is either an integer or a list whose elements may also be integers or other lists. may also be integers or other lists.
The standard solution is recursion. Define a function, passing in the current list and depth:
- If an integer is encountered, the
val * depthCumulative; - If a list is encountered, it is called recursively, depth + 1.
I quickly wrote the recursive version, and after the interviewer confirmed that it was correct, he asked a follow-up question: how to optimize the time complexity.
This is where the BFS The way to do it, layer order traversal, add the integer of each layer to the result first, and then pass to the next layer. This avoids double-counting depth and the code logic is clearer.
The difficulty of this question is not in recursion, but in thinking about follow-up, can you quickly switch ideas. A lot of people only write DFS, but they don't realize that BFS can also be solved. The interviewer is mainly looking at your problem solving ability.
Rhetorical questioning session
Finally went into a rhetorical question. I asked about the requirements related to the position and the interviewer shared some internal LinkedIn:
- Application engineers need to become proficient with AI tools to improve their daily development efficiency;
- Focus more on design capability and impact, not just on how many lines of code you write;
- Communication with the PM is very important, and many design decisions need to be aligned with the product.
The overall atmosphere is relatively relaxed, the interviewer is also very serious when answering questions, and I feel that this company is still quite focused on the growth of engineers.
Interview summary
The overall experience of this round was pretty good, with questions at Leetcode Medium level of difficulty:
- The first question examines the transformation of number theory thinking;
- The second question examines recursion + follow-up ideas.
The real challenge is Integrity of thought: you not only have to write the code, but you also have to be able to explain the complexity and consider optimizations. Project sessions are also not just a casual conversation, but will go deeper and ask for details.
It's really important to mock in advance to help you familiarize yourself with the rhythm of the interview, especially the self-introduction and expression of ideas. If you only organize your thoughts on the spot, it's easy to speak in a messy way.
Getting a big offer? Programhelp
Programhelp's mock and real-time reminders were a big help in getting the answers right this time:
- Remote mock simulation: Rehearsed the full process in advance, practicing the common questions and the sequence of ideas.
- Voice Assist TipWhen you're stuck in an interview, a real-time reminder to "think in terms of the number of factors" or "switch to BFS" can be an instant wake-up call.
- Code commenting and complexity analysis: Not only write the code, but also help you add detailed comments and complexity analysis to make the answer more complete.
For VOs like LinkedIn, coding itself is not the most difficult part, the key is logic and thinking. Our programhelp has already helped many students successfully get VOs at LinkedIn, Amazon, and Google. If you're preparing for similar interviews, you can consider looking for us to assist you, as the efficiency will be greatly improved.
