As a tech giant, Meta is a highly competitive place to work, with thousands of candidates flocking to it every year in the hopes of landing a dream job. In this talent war, the Meta Online Assessment (OA) is the first technical hurdle job seekers face. It’s a timed, high-pressure coding challenge designed to filter out the top technical talent from a sea of resumes, paving the way for future interview rounds.
Join the OAassist interview services team as we dive into the details of the Meta OA and VO interviews, exploring what they test and how you can prepare for them effectively.
Meta OA Overview
Meta’s online assessments are typically hosted on platforms like HackerRank or Codility and are primarily for roles like new graduates, interns, and junior Software Engineers (SWE).
Key Information
- Duration: 70-90 minutes
- Languages: Supports popular programming languages like Python, Java, C++, and JavaScript
- Difficulty: Medium to hard
- Environment: Online, with no internet access. Test cases are limited, requiring you to consider all possible edge cases.
- Structure: Usually consists of 2-3 programming questions and 1 systems design question. Sometimes, it also includes a short behavioral question to assess your values.
Technical Focus
- Programming questions: These focus on algorithms, data structures, and code optimization. The difficulty level is on par with medium-to-hard LeetCode problems, requiring a solid foundation in programming.
- Systems design questions: These require you to design scalable systems and make trade-offs between latency, scalability, and reliability.
- Behavioral questions: Short Q&A designed to evaluate your alignment with Meta’s company culture and values.
Sample Meta OA Question: Counting Geometric Progression Subarrays
To help you better understand the style of Meta’s OA questions, let’s look at a classic problem.
You’re analyzing network traffic logs to identify potential security threats. A specific threat is characterized by a set of packet sizes that form a geometric progression with a given common ratio, r.
Given an integer array packetSizes representing the packet sizes captured in a network session and an integer r, calculate and return the number of continuous subarrays in the array whose elements form a geometric progression with a common ratio of r.
Note: You don’t need to provide an optimal solution. A solution with a time complexity no worse than O(packetSizes.length2) will pass.
Examples
1. packetSizes = [2, 6, 18, 54, 108], r=3
Output: 7
Explanation: The original explanation seems to be mistaken. The actual geometric progression subarrays that satisfy the conditions should be:
- Subarrays of length 1: [2], [6], [18], [54], [108] (5 total)
- Subarrays of length 2: [2, 6], [6, 18], [18, 54], [54, 108] (4 total)
- Subarrays of length 3: [2, 6, 18], [6, 18, 54], [18, 54, 108] (3 total)
- Subarrays of length 4: [2, 6, 18, 54], [6, 18, 54, 108] (2 total)
- Subarrays of length 5: [2, 6, 18, 54, 108] (1 total)
The total count should be 5+4+3+2+1=15.
The question may intend to count geometric progressions with a length greater than or equal to 2, which would result in a total of 4+3+2+1=10. Another possibility is that the output of 7 relates to a different definition or an error in the sample. A common understanding in LeetCode-style problems is that subarrays of length 1 are considered geometric progressions. The discrepancy between the provided example output of 7 and our calculations suggests that the problem’s specific definition might differ from our usual assumptions. This highlights the need to pay close attention to the details of the problem description in a high-stakes test like the Meta OA.
2. packetSizes = [5, 5, 5, 5], r=1
Output: 10
Explanation: With a common ratio of 1, every subarray is a geometric progression. There are 4 subarrays of length 1, 3 of length 2, 2 of length 3, and 1 of length 4. The total is 4+3+2+1=10.
Input/Output Format
- Time Limit: 3 seconds (Java)
- Memory Limit: 1 GB
- Input: array.integer packetSizes An integer array representing network packet sizes.
The Meta Interview Process
Resume submission: The waiting period from submitting your resume to receiving an interview invitation can be anywhere from three weeks to three months. It’s worth noting that if a specific job closes, there’s a “cooling-off period” during which you cannot reapply for the same type of position.
Phone screen: An HR representative will call to discuss your visa status, professional background, and availability for interviews. During this stage, it’s important to be confident and friendly, showing the HR team your passion for and genuine interest in the job. This stage usually does not include a technical interview!
Online assessment (OA): The OA is a moderately difficult assessment that needs to be completed within 70 minutes and consists of 4 questions. Typically, HR will send you the link on the same or next day, and it will be valid for one week. This is where you get to show off your skills! Another possibility is that you might be invited to a phone interview instead.
Phone interview: An engineer will chat with you and might ask two unrelated questions or one question with follow-ups. The difficulty of the algorithm questions is the same for both interns and new graduates.
Virtual onsite (VO): The HR team usually schedules the virtual onsite interviews for two weeks later. This could include three coding rounds and one behavioral round, or one systems design round. Starting in 2023, it’s possible to request with HR to split the interview into two days, whereas before, it was a single-day event.
VO Interview Question
Given a string containing only lowercase letters and the # symbol, decode it back to the original string. The encoding rules are: a-i correspond to the digits 1-9, and j-z correspond to the digits 10#-26#.
Examples:
- The string “10#11#12” should be decoded to “jkab”.
- The string “1326#” should be decoded to “mzf”.
At first glance, this interview question seems like a simple string parsing problem, but it’s a test of your understanding of greedy algorithms. Since the digit 1 can either represent the letter a or be part of a two-digit code like 10# for the letter j, the processing order becomes critical.
The key to solving this problem is to iterate through the string backward. By starting from the end, whenever you encounter a #, you know it must be part of a two-digit code with the two digits preceding it (e.g., 10#). If you don’t encounter a # at your current position, you can confidently decode the single digit. This backward-processing approach ensures you always make the optimal choice, avoiding ambiguity.
Meta VO Systems Design
Your task is to design a video-on-demand platform similar to Netflix or YouTube.
The interviewer makes it clear that the focus isn’t on the nitty-gritty, low-level technical details like specific codecs or CDN strategies. Instead, they want you to think like a product manager, focusing on the user experience and system functionality. The key is to architect a system that runs smoothly and provides a comfortable experience for users.
Here are the requirements:
- How do you handle video uploads? When a user uploads a high-definition movie, how does the system ensure the file is processed quickly and securely, and how does it handle necessary transcoding?
- How do you implement a video recommendation feature? When a user opens the homepage, how do you accurately push videos they might be interested in, based on their preferences and viewing history?
- How do you handle sudden traffic spikes? When a video goes viral and attracts millions of views in a short period, how does the system handle this massive influx of traffic?
- How do you optimize the video playback experience? How can you ensure users enjoy a smooth, buffer-free playback experience regardless of their location or device?
- How do you design a commenting and liking system? How do you ensure users can post comments and like videos instantly, and that this interactive data is updated in real time for all users to see?
The problems listed above are some of the common interview questions OAassist has collected and compiled while assisting our clients. Of course, Meta’s interview difficulty goes far beyond these examples. With the rise of AI, Meta has significantly raised its hiring bar, making its interviews arguably the most difficult in the current Silicon Valley job market. While Meta offers some of the best salaries and benefits among the FAANG companies, simply grinding LeetCode problems isn’t enough to guarantee a pass.
If you’re preparing for a technical role at Meta, consider reaching out to OAassist for VO interview assistance or proxy services to improve your chances of success without the tedious grind of repetitive practice. We wish you the best of luck in your job search!
