J.P. Morgan Software Engineer II Interview Experience — Four Rounds from System Design to RAG

Round 1: Hiring Manager

System design: choose a project you previously worked on, design and explain its architecture end to end, and answer related follow-up questions.

Database optimization: given a database table with user_id and region columns and more than one million rows, how would you optimize a query that looks up a specific record?

1. Database indexing.
2. Table partitioning.
3. Sharding.
4. Replication.
5. Discuss the trade-offs among those methods.

Python fundamentals and networking concepts:

1. Multithreading versus multiprocessing.
2. Differences among PUT, POST, and PATCH.
3. How garbage collection works.
4. How Python handles garbage collection.
5. Deadlocks.

DSA: explain the basic logic of Kadane's algorithm.

Round 2: VP Round

Technical fundamentals:

1. Parallelism versus concurrency.
2. Practical use cases for a stack, including about ten minutes of deep discussion and several follow-ups.
3. Mutable versus immutable data types.

Logic: two logical puzzles.

Managerial behavioral questions:

1. Why do you want to change jobs?
2. Have you discussed your decision to leave with your current manager?

Round 3: Senior VP Round

AI, machine learning, and RAG questions:

1. What is the difference between a skill and an agent?
2. What are LangChain and LangGraph?
3. Give a deep explanation of how a RAG architecture works.
4. What different types of embeddings are there?
5. How do you perform chunking?
6. How are chunks stored in a database?
7. What preprocessing steps are needed for images and text?
8. What are the differences among accuracy, precision, and recall?
9. In a RAG architecture, how do you retrieve data from the database?

Systems and security design questions:

1. How would you send a large file to another endpoint?
2. What is the difference between authentication and authorization?
3. Can you create a role that restricts a user from performing certain operations with a token?
4. How does token-based authorization work?
5. What does the architecture of a token-based authorization system look like?

Round 4: VP Round

Python and systems fundamentals:

1. Can a list be stored inside a tuple?
2. How does memory cleanup work?
3. Explain the CAP theorem.
4. How does scaling work in Kubernetes?

DSA problems:

1. Solve the K largest integers problem and explain in detail how the elements are arranged at each step of the algorithm.
2. Explain the logic of merge sort.

Work-experience follow-ups: detailed questions about prior work experience and the architecture of systems that had been developed.