I just finished helping a trainee walk through a set Google SDE Virtual Onsite, the whole process from e-interview to manager interview is quite interesting.Google's interview style is still the same as before: coding, plus system design and communication skills. Algorithm is the foundation, but more important is the ability to express your thoughts, system design, and whether you can optimize your answers under the guidance of the interviewer. Here's a breakdown of the complete process
Round 1: Telephone Interview (Coding)
The phone interview was 45 minutes long and consisted of the most common algorithmic questions. This time, the question was about Dynamic Programming (DP), which is similar to a variant of "Yang Hui Triangle".
The trainee started off naturally with recursion, and after writing it himself, he realized that the time complexity wasn't quite right. The interviewer followed up by asking, "Can you optimize it?" So he immediately switched to iterative DP, using an array to store intermediate states, and the complexity dropped from exponential to O(n²). In the end, the test cases all passed.
Summary:
Google electric interview is actually more like "qualification screening", will not give particularly difficult questions;
The focus is on the ability to quickly grab the state transfer formula for DP and the handling of boundary conditions;
It would be a big plus to volunteer complexity and optimization points.
Round 2: Programming Interview (Coding)
This round was a video interview, the interviewer was an Indian with a heavy accent, and the trainee had to double check the meaning of the questions at the beginning. The topic was the classic Binary Tree Zigzag Traversal (Binary Tree Hierarchical Traversal variant).
The trainee started out on the wrong path: first wrote a middle-order traversal, and found that the results were completely wrong; then tried the back-order, or wrong; finally, in the interviewer hint, only to return to the BFS + queue of the right way, write and pass.
Although he has taken a wrong turn here, the interviewer is actually very concerned about the process of revising his thinking: can he listen to feedback? Can you be flexible and adaptable? This is more important than just getting it right.
Summary:
Google values a candidate's resilience to solve problems;
If you get stuck, don't take it on the chin, talk to the interviewer directly and admit your mistakes;
BFS, DFS, and tree traversal are high-frequency test points for Google coding, so you really need to brush up on them in advance.
Round 3: Coding + Ideas Explained
In the third round, the level of difficulty increased, not only to write code, but also to write and talk about it, and to discuss time and space complexity, and to be able to propose optimization solutions.
The trainee first writes a solution for O(n²) and then adds to himself, "Can this be optimized to O(n)?" The interviewer asked him about it, and he adjusted his thinking on the spot and gave the optimized implementation. The whole process demonstrates the evolution of thinking, the effect is very good.
Summary:
In the Google interview, it's okay if the first version of the code isn't optimal, but it's important to actively think about whether there is a better solution;
The interviewer wants to see you go from naive → optimized, not just give a rigid answer;
Do not be silent for too long, and verbalize your thoughts even if you haven't thought them through yet.
Round 4: Manager Interview (OOD + System Design)
The final round is an on-site interview led by the manager, focusing on Object Oriented Design (OOD) + system design skills. The topic was e-commerce scenario:
search function:
Design the index structure so that users can search for items quickly;
The results should support sorting and filtering;
Consider also paginated displays.
Shopping Cart & Checkout Process:
Basic functions (adding, deleting, and checking products, submitting orders);
Exceptions (insufficient stock, payment failure);
Extension points (referral system, coupon logic, secure payment links).
The trainee did very well in this round, not only explaining the technical architecture clearly, but also adding some thoughts from the perspective of user experience, such as: to give a friendly hint when the payment fails, how to ensure the security when checking out. This makes Manager feel that he is not just a "code writer", but an engineer who can think from the product perspective.
Summary:
Google system design questions are not about whether you can write them or not, but about whether you can break them down and abstract the modules;
Trade-off analytics are a plus, such as why caching? Why microservices?
Speaking in terms of user experience often makes a better impression on the Manager.
Participants' feelings
After coming out of the student said a representative sentence: "Google's questions are not as difficult as the legend so scary, but the whole process will force you to keep explaining and optimizing. Writing correctly is just the foundation, what is more important is to let the interviewer see your 'thinking and growing process'."
Google SDE routine dismantling
The Google SDE VO set is probably:
Coding: DP, trees, and BFS/DFS are high-frequency exams;
train of thought: Code is not the point, being able to speak clearly is;
System Design: Ability to abstract into modules from users to architecture;
Communicate and collaborate: Courage to admit mistakes and adjust direction in a timely manner.
If you are also preparing for Google SDE VO recently, you really don't want to brush up on the questions by yourself ~ a lot of details can only be stepped out during the simulation. We have remote voice reminders and interview aids that can help you wake up in time when you are stuck, so you won't be pressed for time. It's much more efficient to practice together than to work alone!
