Comprehensive Big Data Interview Questions and Preparation Guide for Campus Graduates

Today’s protagonist is a fresh graduate who, despite a slowing economy and a saturated internet market, has secured offers from major internet firms such as Tencent, Ant Group, and Bilibili.

The author, acting as a mentor, poses several questions from a job‑seeker’s perspective and provides answers available in a linked Knowledge Planet module.

1. What experiences do campus recruiters value most?

Beyond the standard algorithm and “八股” (rote) questions, interviewers look for deep project experience. Self‑built projects that demonstrate real engineering challenges are far more compelling than generic or toy projects.

2. Summarized interview questions asked by various companies

Below are the collected questions:

Bilibili

1. Explain the GC model
2. Have you performed GC tuning?
3. What do you know about YARN (describe the submission process)
4. How does HQL translate to MapReduce source code?
5. HBase read/write flow
6. Hive data skew
7. Discuss your internship, project, and architecture design
8. Scenario: In Hadoop, RM schedules resources – which resources would you aggregate and how?
9. Scenario: After aggregation, data volume is huge, RM pressure is high – how to resolve?
10. Algorithm: QuickSort

1. Code performance issue – how to locate the bottleneck?
2. Does Java manage off‑heap memory?
3. Difference and similarity between page cache and buffer (pros/cons, use cases, Linux 2.4+ unification)
4. How do you schedule your time?
5. Recent big‑data ecosystem features you follow?

ByteDance

1. What did you do during your internship?
2. How to minimize downstream impact of upstream data changes?
3. Explain checkpoint (aligned vs. unaligned, incremental, 1.16 generic incremental ck)
4. If a Flink job’s topology changes (e.g., new source), how to reuse previous checkpoint?
5. State backend
6. How does Flink achieve exactly‑once?
7. Flink minibatch
8. Broadcast variables
9. Kafka isolation levels
10. Hive internal vs. external tables
11. Algorithm: Implement queue with two stacks

Ant Group

1. HQL to MR process
2. Flink exactly‑once implementation
3. Internship experience
4. Explain HashMap
5. HashMap put() source code
6. equals vs. hashCode
7. How to ensure thread safety for i++ (atomicity)
8. Similarities and differences between synchronized and ReentrantLock
9. Thread pool’s seven parameters
10. Java class loading mechanism
11. CMS tuning parameters
12. Do you know the JVM start‑up memory size parameter?
13. How to locate OOM source code
14. MySQL left‑most matching principle
15. MySQL index structure
16. How to optimize indexes
17. HTTP vs. HTTPS
18. How user login is identified
19. How to share sessions across multiple servers
20. Can a stolen cookie bypass normal login?
21. Algorithm: LRU

Tencent

1. How to store a 10 GB file in HDFS?
2. Your understanding of Raft and Zookeeper
3. Explain Kafka exactly‑once implementation (detail the 2PC process)
4. Difference between process, thread, and coroutine
5. Java Memory Model
6. Explain GC partitioning
7. How to decide which objects need reclamation
8. Why is LSM‑tree suitable for massive writes? Why not other structures?
9. LSM‑tree compaction process

Most interviews start with a project discussion, then dive into framework principles and “八股” questions, exploring business pain points and technical fundamentals.

3. How to participate in open‑source community activities?

For students unable to secure offline internships, contributing to open‑source projects offers location‑independent experience and strengthens resumes. Contributing to Apache projects showcases coding ability, lets interviewers review your PRs, and demonstrates deep understanding of real‑world problems solved by the community.

Final Summary

The campus recruitment landscape has become significantly tougher; some questions now challenge even engineers with three to five years of experience. The industry is moving toward two trends: elite positions that command high salaries and outsourced or internal‑staffed roles that require less specialized expertise.

Candidates should seize the remaining window to transition from smaller firms or outsourcing positions to core companies, continuously learn, and seek breakthrough opportunities.

Regardless of background, job seekers must adapt their mindset to a stabilizing or declining market, recognizing that salaries in other industries may be three times higher, adding pressure to stay competitive.