I recently navigated the onsite interview process at Facebook, and I've detailed the specific LeetCode problems I encountered during my rounds below. Having also interviewed with Google, Amazon, and Apple, I've developed a perspective on the relative difficulty of securing a position at each. In my experience, Facebook felt somewhat more accessible than Google, particularly if one has thoroughly prepared using the Facebook-tagged questions on LeetCode; I personally didn't encounter any questions that weren't already on LeetCode. On a difficulty scale, I'd rate Google around a 9, Facebook a 7 (or even a 5 if you've mastered their tagged questions), Apple a 5, and Amazon a 2 or 3, as it primarily focused on fundamental concepts. Although this post recounts my Facebook experience, I ultimately accepted an offer from Google, and things have been going great there. To anyone currently interviewing, keep up the excellent work – you'll achieve your goals!

Last week, I had my virtual onsite interview with facebook for the Sr Software Engineer (E5) position. The overall experience was great.

Phone Screen

This was a 45-minute round where the interviewer asked two coding questions. I focused on understanding the problem, clarifying any ambiguities, and then verbally explaining my solution along with its time and space complexity. Once the interviewer was satisfied, I quickly implemented the solution and performed a dry run with test cases, including corner cases, finally explaining the complexity of my written code.

Virtual Onsite

The virtual onsite consisted of four 45-minute rounds: two Coding Rounds (Ninja), one System Design Round (Pirate), and one Behavioral Round (Jedi).

Coding Round 1:

Similar to the phone screen, I was given two coding problems to solve, applying my structured approach of understanding, verbalizing, implementing, and verifying.

System Design Round:

I was asked to design the privacy settings at facebook. I spent approximately 40 minutes discussing the design, and used the remaining 5 minutes to ask questions about facebook's culture.

Coding Round 2:

This round also involved two coding questions, where I applied my standard approach for solving them.

Behavioral Round:

The interviewer started by asking me about myself, my current project, and significant accomplishments. I was also asked if I had ever managed conflicting priorities and if I had received constructive feedback from a colleague. For each behavioral question, I provided a supporting anecdote from my past experiences, ensuring my answers aligned with facebook's mission and core values. Interestingly, this round also included one coding question.

Two days after my virtual onsite, I received a call from the recruiter informing me that the feedback from my interviewers was good, and they are still waiting for the hiring committee to make the final decision.

Hey Everyone,

I interviewed with Facebook for an E4 role in London. I am currently waiting to hear back from them regarding the outcome of my interviews.

Phone Screen:

This round involved two coding questions. I felt confident in my ability to solve both on time.

1. Binary Tree to Doubly linked List.
2. Evaluating expression (Only with +/* operators).

Onsite Coding 1:

I successfully solved both problems presented in this round within the given time frame.

1. Multiplying 2 strings.
2. Course Schedule problem on LeetCode.

Onsite Coding 2:

Only one question was asked in this round. I was able to solve it on time, even though I required one hint from the interviewer. Overall, the interviewer seemed convinced by my solution.

1. LFU Cache Implementation.

Behavioral:

This round consisted of standard behavioral questions and some inquiries about my past work experience. I managed to tackle all the questions well, with only one or two proving to be a bit tricky.

System Design:

The challenge here was to design a system similar to Google Drive. I was able to approach the problem, but the interviewer continuously introduced new scenarios. I could answer some of these, but got stuck on others. I feel this round went below average to average.

I had an interview for an E5 Lead Engineer position at Facebook in London. My experience level was 12 years.

Talk with the recruiter (30 mins):

This discussion was crucial for setting my expectations and understanding the interview process. The recruiter provided valuable insights and resources for preparation. I treated this as an interview itself, asking all my questions about the preparation process.

Technical Phone Screen (45 mins):

I was asked two coding questions in this round. Both were relatively easy, and the key was to articulate my thought process clearly. I made sure to test my code immediately after writing it, which impressed the interviewer.

Virtual Onsite (4 rounds):

Behavioral (45 mins):

This round focused on my past experiences in teams and how I handled technical challenges. The questions were open-ended, and each answer led to follow-up questions. I wasn't asked any coding questions in this round. My recruiter had mentioned that a strong positive signal in this round is critical, as a negative signal here can lead to a 'no-hire' decision even with strong technical performance.

System/Product Design (45 mins):

I was tasked with designing a file-sharing system similar to Dropbox. During requirement gathering, I pointed out a potential security issue with auto-sync, which seemed to impress the interviewer. My proposal for staggered scale-up based on usage trends also received positive feedback. I drove the discussion, answering clarification questions and providing high-level differences between SQL and NoSQL without getting bogged down in details.

Coding 1 (45 mins):

This round was similar to the phone screen but with a higher bar. I tackled two coding problems. The first involved a LeetCode problem, where I used Breadth-First Traversal and discussed follow-up questions on parallelization and memory constraints. The second was another LeetCode problem, which I solved very quickly.

Coding 2 (45 mins):

I faced two more coding questions here. For the first, which involved checking interval coverage, I initially missed a boundary case but caught and fixed it during testing without prompting, which was viewed positively. For the second, finding connected sub-graphs in an undirected graph, I used Depth-First Traversal. I ran out of time to complete the code, but because I had modularized my solution, I could clearly outline the remaining implementation details for the interviewer.

The next day, I received positive feedback from my recruiter for all four rounds and was informed that I would proceed to the review board. The offer came through the day after that.

I had my Facebook E5 phone screen interview, which lasted for 45 minutes. We started with a brief 3-5 minute introduction. The interviewer then clarified that the majority of the time would be dedicated to problem-solving, reserving the last 5 minutes for my questions.

The first question I encountered was Add Binary. Having practiced this problem before, I immediately explained the optimal approach, achieving a time complexity of O(max(l1, l2)). The interviewer was satisfied with my algorithm, and I quickly proceeded to code the solution in Java. After coding, I walked through my solution with a few examples. This entire segment, including discussion and coding, was completed within 10-15 minutes.

/**
 * Time Complexity: O(max(l1, l2))
 * Space Complexity: O(max(l1, l2))
 * l1 = Length of string A, l2 = Length of string B.
 */
public String addBinary(String a, String b) {
    if (a == null && b == null) {
        return "0";
    }
StringBuilder result = new StringBuilder();

int i = a.length() - 1;
int j = b.length() - 1;
int carry = 0;

while (i >= 0 || j >= 0) {
    int sum = carry;
    if (i >= 0) {
        sum += a.charAt(i) - '0';
        i--;
    }
    if (j >= 0) {
        sum += b.charAt(j) - '0';
        j--;
    }
    result.append(sum % 2);
    carry = sum / 2;
}

if (carry == 1) {
    result.append('1');
}

return result.reverse().toString();

}

Following that, the second question was Merge Intervals. Again, I was familiar with this problem and presented the optimal O(NlogN) approach. Once the interviewer approved the algorithm, I implemented the solution in Java and demonstrated its functionality with test cases. This question took approximately 20-25 minutes to complete.

/**
 * Time Complexity O(N log N)
 * Space Complexity: O(N) (Including the List<int[]> result)
 * N = Length of input array.
 */
public int[][] merge(int[][] intervals) {
    if (intervals == null || intervals.length <= 1) {
        return intervals;
    }
Arrays.sort(intervals, (a, b) -&gt; (a[0] - b[0]));

List&lt;int[]&gt; result = new ArrayList&lt;&gt;();
int start = intervals[0][0];
int end = intervals[0][1];

for (int i = 1; i &lt; intervals.length; i++) {
    if (intervals[i][0] &lt;= end) {
        end = Math.max(end, intervals[i][1]);
    } else {
        result.add(new int[] { start, end });
        start = intervals[i][0];
        end = intervals[i][1];
    }
}

result.add(new int[] { start, end });
return result.toArray(new int[result.size()][2]);

}

With a few minutes remaining, I took the opportunity to ask the interviewer two questions about their current work at Facebook. I was thrilled to later receive a call from the recruiter informing me of positive feedback from the hiring committee and that an offer would be extended.

I recently had a virtual onsite interview with Facebook for an E5 Software Engineer role in the Bay Area. The process involved four 45-minute rounds: two focused on coding, one on system design, and a behavioral round which surprisingly also included a small coding problem.

1st Coding Interview

The first coding round started with a problem similar to Intersection of Two Linked Lists, but presented in the context of a binary tree. I was asked to find the Lowest Common Ancestor (LCA) of two nodes in a binary tree, where each node only had a parent pointer. Initially, it took me a moment to connect it to the linked list intersection concept, but I eventually figured it out without any hints. I implemented a two-pointer solution in Java, achieving O(n) time complexity and O(1) space complexity. After walking through my code with a few examples, the interviewer seemed pleased. We then moved to a second coding question: Leftmost Column with at Least a One, which was identical to the problem found on LeetCode. I provided an optimal solution for this as well.

2nd Coding Round

In my second coding round, I tackled two problems. The first was Range Sum of BST, and the second was Trapping Rain Water. Both were exactly as they appear on LeetCode. I developed optimal solutions in Java for both. For the first question, the interviewer requested a few examples and base cases to test my solution. Unfortunately, we ran out of time before I could thoroughly walk through the solution for the second problem.

System Design Interview

The system design interview challenged me to design a Proximity Server. The core requirement was to, given a location's latitude and longitude, find all nearby places within a specified radius. I began by clarifying the essential features and discussing the design goals and potential scale. I then outlined a high-level design of the APIs. A significant portion of my discussion focused on the data structures and algorithms for efficiently finding nearby places; I proposed using a Quad Tree to represent geographical grids and store locations. I also delved into sharding strategies, briefly explaining Consistent Hashing, and the importance of caching for quick data retrieval. Additionally, I touched upon the application layer, mentioning stateless machines and load balancers. I later found a helpful video on this topic here.

Behavior Round

The behavioral round primarily centered on my projects at my current company. The interviewer also posed questions related to conflict resolution, how I've made an impact outside my immediate team, and my approach to problem-solving when faced with limited information. For each of these, I drew upon my past experiences to provide concrete examples. Towards the end, a simple coding problem was introduced: Valid Palindrome. I believe I performed well in this round, effectively communicating my experiences and solving the coding challenge.

Overall, I feel the interview went quite well. I've received an email from the recruiter indicating that it might take up to a week to hear back about the result. So, I'm keeping my fingers crossed!