ACCELDATA – Sr SDET Interview (Round 2)

Questions & Sample Answers

Overview: These were the actual questions asked to me in the Acceldata Senior SDET (Round 2) interview.

My experience level at the time of the interview was 4+ years.

This round focused heavily on:

Framework architecture
Concurrency & rate limiting
Distributed system testing
Kubernetes & Big Data scenarios
Real-world debugging

The discussion tested both depth and structured thinking.

Question 1: Tell me briefly about yourself.

Answer:

I have 4+ years of experience in QA and Automation, primarily focused on API automation using Python and Pytest.

I have worked in microservices-based systems and have experience integrating automation into CI/CD pipelines.

I also have exposure to Docker, Kubernetes, and cloud environments, along with hands-on debugging of production-like issues.

Question 2: Explain your Tech Stack and Roles & Responsibilities.

Answer:

Tech Stack:

Python, Pytest
REST APIs
Jenkins
Docker
Kubernetes (basic exposure)
SQL
AWS exposure

Roles & Responsibilities:

Requirement analysis
Automation framework development
CI/CD integration
Debugging test failures
DB validation
Handling production issues
Cross-team collaboration

Question 3: Explain your automation framework.

Answer:

My framework is modular and scalable.

It includes:

Reusable API client layer
Utilities and helper modules
Centralized configuration management
Logging and reporting support
Environment-based execution
CI/CD integration

The design follows separation of concerns to ensure maintainability and scalability across microservices.

Question 4: What technical challenges did you face while building the framework?

Answer:

Some challenges included:

Managing environment-specific configurations
Handling dynamic test data
Managing microservice dependencies
Flaky tests due to timing issues
Rate limiting during parallel execution
Standardizing logging across modules

These were resolved using abstraction, retry mechanisms, proper fixture design, and structured logging.

Question 5: What is the difference between 401 and 403?

Answer:

401 – Unauthorized

Authentication failed
Token missing or invalid

403 – Forbidden

Authentication succeeded
But user does not have permission

Question 6: Are you aware of rate limiting? How do you handle 200 requests/min limit when tests run in parallel?

Answer:

Yes.

If there is a limit of 200 requests per minute:

Handling strategies include:

Implement retry mechanism with exponential backoff
Introduce request throttling
Use shared rate limiter logic
Reduce parallel threads
Mock dependent APIs if possible

The goal is to avoid overwhelming the service and reduce flakiness.

Question 7: Have you faced connection pool exceptions?

Answer:

Yes.

This usually happens due to:

Too many concurrent connections
Improper connection closing
Small connection pool size

Handling includes:

Increasing pool size appropriately
Ensuring proper connection closure
Controlling parallel execution

Question 8: Have you faced concurrency exceptions? How did you handle them?

Answer:

Yes, especially during parallel test execution.

Common causes:

Shared test data
DB write conflicts
Shared mutable state

Handling strategies:

Generate unique test data
Avoid shared state
Use transaction isolation
Proper synchronization where required

Question 9: Merge two sorted arrays into one sorted array without duplicates. Explain time and space complexity.

Answer:

Approach:

Use two pointers
Compare elements
Add smaller element to result
Skip duplicates

Time Complexity:
O(n + m)

Space Complexity:
O(n + m)

Question 10: How would you test a Spark ETL pipeline running on Kubernetes (HDFS → ADLS)?

Answer:

High-level testing approach:

1. Infrastructure Readiness

Validate Kubernetes cluster health
Verify Spark pods scheduling
Validate storage connectivity
Check IAM and permissions

2. Application Readiness

Verify Spark job configuration
Validate environment variables
Validate resource allocation

3. Data Validation

Validate input data format
Schema validation
Record count validation
Transformation logic validation
Output data consistency

4. End-to-End Flow

Trigger job
Monitor logs
Validate job completion
Validate output storage

Question 11: Before processing starts, how do you validate complete readiness?

Answer:

Validate infra setup (K8s, storage, networking)
Validate application configuration
Validate permissions and secrets
Validate resource allocation
Validate connectivity between components
Run a smoke job before full execution

Question 12: Data store is on different machine and Kubernetes is running in EKS. How will they communicate?

Answer:

Communication through:

VPC configuration
Security groups
Network routing
DNS resolution
Exposed service endpoints

Ensure ports and firewall rules allow communication.

Question 13: Have you worked with Big Data / Hadoop?

Answer:

I have basic exposure.

Understanding includes:

HDFS concepts
Distributed data storage
Spark job flow
ETL pipelines

My focus has been more on testing strategy rather than development.

Question 14: Linux commands you have used.

Answer:

ls
cd
pwd
grep
cat
tail -f
ps
top
chmod
kill

Used mainly for debugging and log inspection.

Question 15: kubectl commands you have used.

Answer:

kubectl get pods
kubectl describe pod
kubectl logs
kubectl get services
kubectl get deployments

Used for monitoring and debugging Kubernetes workloads.

Question 16: Why is Kubernetes more reliable than plain Docker deployment?

Answer:

Docker:

Runs containers
Single-host management

Kubernetes:

Self-healing
Auto-scaling
Rolling deployments
Load balancing
Resource management

It provides production-grade orchestration.

Question 17: Are you aware of Docker Compose?

Answer:

Yes.

Docker Compose:

Used for multi-container setup
Defined using docker-compose.yml
Mostly used in local development
Not suitable for large-scale production orchestration

Key Takeaways

Round 2 focuses on architecture maturity.
Concurrency and rate limiting scenarios are important.
Distributed system testing requires layered validation.
Infrastructure awareness differentiates senior candidates.
Clear structured answers create strong impact.