TELL ME ABOUT YOURSELF

My name is Zhibek. I am a DevOps Engineer with more than 5 years of experience in AWS cloud. I started at Yelp in 2021, where I worked 4 years on high traffic consumer apps. I built AWS infrastructure, ran Kubernetes on EKS, and improved our CI CD pipelines with GitHub Actions. With blue green deployments and automatic rollback, I made deployments 35 percent faster. I also mentored junior engineers and wrote runbooks for the team. Now I work at Zulily, where I build Terraform for dev, staging and production, run EKS with autoscaling, and build secure pipelines with approval steps. I have AWS Solutions Architect Associate and CKAD certifications. Now I am ready for the next step, to lead a team as a player coach, and that is why this role at Herc Rentals is very interesting for me.

DAY TO DAY RESPONSIBILITIES

My day starts with checking CloudWatch dashboards and alerts to see if anything broke at night. Then we have a standup with the team. Most of the day I work on Terraform changes, like new environments, network changes or new services. I review pull requests from other engineers. I help developers when their pipeline fails or their app does not scale well on EKS. I fix production issues when they happen and do root cause analysis after. I also update documents and runbooks so the team knows how to handle things. Sometimes I meet with security or app teams to plan new projects.

TELL ME ABOUT YOUR PROJECT

At Zulily the team needed a private database that is easy to repeat in all environments. I wrote Terraform modules for Amazon DocumentDB, put it in private subnets, and used security groups and IAM so only approved apps on EKS can connect. I added it to GitHub Actions with plan, review and approval before production. I set CloudWatch alarms for CPU, connections and storage. The result was the same setup in dev, staging and production, no public access, and new environments now are built much faster.

Second project if they ask, at Yelp. Deployments were slow and risky. I optimized GitHub Actions with caching and parallel jobs, added blue green deployments, and added automatic rollback if health checks fail. Deployments became 35 percent faster and we had fewer failed releases.

TECHNICAL QUESTIONS AND SHORT ANSWERS

What is Terraform state and why remote state
State is a file where Terraform keeps what it created. I keep it in S3 so the whole team uses the same state, and DynamoDB locks it so two people cannot change at the same time.

How do you structure Terraform for many environments
I use reusable modules and separate folders or workspaces for dev, staging and production, with different variable files. Same code, different values.

What is Terraform drift and how do you handle it
Drift is when someone changes things by hand in the console. I run terraform plan regularly to find it, and then fix it in code, not by hand.

Terraform vs CloudFormation vs Bicep
CloudFormation is only AWS, Bicep is only Azure, Terraform works on many clouds. For a multi cloud company like Herc, Terraform is the best choice for one standard.

What is a landing zone
It is a ready secure base for cloud accounts, with account structure, network, IAM, logging and guardrails. In AWS it is Control Tower and Organizations. New teams get a safe account fast.

How does Kubernetes autoscaling work
HPA adds more pods when CPU or memory is high. Cluster Autoscaler adds more nodes when pods cannot fit. Together they handle traffic changes.

A pod keeps crashing, what do you do
I run kubectl describe pod and kubectl logs. Common reasons are wrong config, missing secret, out of memory, or failing health checks. Then I fix and redeploy.

What is Helm
Helm is a package manager for Kubernetes. It uses templates and values files so we deploy the same app to many environments in a standard way.

Blue green vs canary
Blue green has two full environments and you switch all traffic at once, easy rollback. Canary sends small traffic first to the new version, then more if it is healthy.

How do you design a CI CD pipeline
Build, unit tests, security scan, terraform plan, approval for production, deploy, health check, and automatic rollback if it fails.

How do you keep secrets safe in pipelines
Never put secrets in code. Use AWS Secrets Manager or Parameter Store, and GitHub OIDC to get short time AWS credentials instead of long keys.

Explain VPC design
Public subnets for load balancers, private subnets for apps and databases, NAT gateway for outbound internet, security groups to control traffic, many availability zones for high availability.

Security group vs NACL
Security group works on instance level and is stateful. NACL works on subnet level and is stateless, so you need rules both ways.

What is hybrid connectivity
Connecting on premise data center to the cloud with Site to Site VPN or Direct Connect in AWS, ExpressRoute in Azure, FastConnect in OCI. Transit Gateway helps connect many VPCs.

What is IAM least privilege
Give only the permissions needed, nothing more. Use roles, not users with long keys, and review permissions often.

How do you do disaster recovery
First know RTO and RPO, how fast to recover and how much data you can lose. Then backups, multi AZ, cross region copies, and test the recovery regularly, not only on paper.

How do you reduce cloud cost
Right size instances, autoscaling, Savings Plans or Reserved Instances, spot for non critical work, delete unused resources, tagging for each team, and monthly cost reports to leadership.

How do you handle a production outage
First stop the impact, maybe rollback. Communicate to stakeholders. Find root cause with logs and metrics. After that write a postmortem with no blame and fix so it does not happen again.

Do you have Azure or OCI experience
Be honest. My deep experience is AWS, but the concepts are the same. VPC is VNet, IAM is Entra ID and RBAC, EKS is AKS or OKE. Terraform works on all of them, so I can learn fast. I already started learning Azure basics.

How do you use AI in your work
I use Claude AI to help draft pipelines, review Terraform and write documentation. I always review and test before using it. I think AI can help the team move faster when used carefully.

MANAGER QUESTIONS

How do you handle a low performer
Talk early in one on one, understand the reason, set clear goals with time, give support, and follow up. If no change, work with HR.

How do you prioritize when everything is urgent
Production issues and security first, then business priorities agreed with leadership. I make the work visible so everyone sees the tradeoffs.

How do you manage contractors
Clear scope, same standards and code review as full time, and documentation so knowledge stays in the company.

Two engineers disagree on a design, what do you do
Let both explain, compare by facts like cost, risk and simplicity, maybe do a small test, then decide and move forward together.

What will you do in first 90 days
First 30 days listen and learn the team, systems and pain points. Next 30 days find quick wins like standards and runbooks. Last 30 days make a roadmap with leadership.

How do you measure team success
Deployment frequency, failed changes, time to recover, uptime, cloud cost, and team happiness.