# NVIDIA Interview, Cloud Storage

- 10/5: Applied online
- 10/16: Received an HR email for interview invitation
- 10/18: Two rounds of 45-minute interviews on the same day
- 10/19: Received an HR offer call

After I told HR about my offer deadline with ByteDance, they kindly arranged two urgent interviews for me and issued the offer the day before the deadline.

The positions might have been unique (leaning towards SRE, not a typical SDE), so neither round of interviews involved Leetcode questions; they focused solely on my resume and relevant technical understanding.

**First Round:**
- Q1: Self-introduction
- Q2: Regarding GCP mentioned in my resume, he asked if I was only familiar with GCP or had experience with other cloud providers:
I answered that I was currently using AWS in my courses and believed the concepts across cloud providers were very interchangeable, so adapting to any of them shouldn't be too difficult.
- Q3: Why did my previous job use Google Kubernetes Engine (GKE), and what problem did it solve:
I responded that the original architecture managed all microservices with docker compose on a single server, which quickly ran into scalability issues. Receiving GCP funding, we discovered k8s as a powerful solution and transitioned the architecture to GKE using kubecompose, successfully resolving the scalability issues.
- Q4: Whether my past experiences were more in software development or system infrastructure:
I mentioned having experience in both as my past teams were small, often requiring me to set up the infrastructure before developing software services. However, if I had to choose, my experience leaned more towards software development, studying system infrastructure out of necessity and immediate need.
- Q5: Besides GKE, which other cloud provider technologies had I used:
I replied with Google Cloud SQL, which we used as a large central MySQL database, and I specifically developed an ORM interface to interact with other services.
- Q6: Some questions on microservices:
I elaborated based on what I wrote in my resume plus some additional details.
- Q7: Whether I built the entire CI/CD pipeline from scratch and what it included:
I affirmed, detailing my creation of the pipeline using Gitlab CI/CD and GCP's image registry, encompassing source control, build & test, integration testing, and production environment setup with alarm and key metrics monitoring.
- Q8: Familiar programming languages:
I mentioned JavaScript, to which he retorted that their team did not use it.
- Q9: Familiarity with the Linux environment:
I explained my experience building a Linux server from scratch and general comfort with its operations, though I was less clear on the distinctions among different Linux distributions. He found my experience sufficient.
- Q10: Questions about K8s resources:
The question was broad; I only covered basics like deployment, service, storage pool, and replica set, along with their functions.
- Q11: Whether I had built a K8s cluster from scratch:
I admitted to only using clusters set up by GCP or those established by others at work, without having built one myself.
- Q12: Courses taken at Columbia University:
I simply responded, and he showed particular interest in cloud computing, prompting further questions about the AWS services used in my assignments.

This concluded the technical questions, followed by two behavioral ones about past mistakes and conflicts with colleagues. I hadn't prepared for these, so I improvised, which wasn't very polished, but he did not comment further, ending with a Q&A session.

**Second Round:**
Similar to the first round but more detailed, such as optimizing image size using docker layer mechanisms. I can provide more details if there's interest.