Resilience_Engineer_Deep_Interview_Flashcards
🔧 TECHNICAL QUESTIONS
How do you decide whether to build a custom solution or use an existing SaaS tool?
S: At Eficens, we had to automate invoice generation for a loan processing system.
T: I had to determine whether to build a custom solution or integrate a SaaS tool.
A: I conducted a cost and feature comparison between building it in-house with Spring Boot and integrating Sage Intacct. I considered factors like supportability, integration time, team bandwidth, and maintenance overhead.
R: We chose Sage Intacct via API integration, reducing dev time by 3 weeks and ensuring compliance. I also automated data exchange using Python scripts for daily sync with PostgreSQL.
Explain a scenario where you integrated legacy systems with new cloud-native solutions.
S: At TCS, our client wanted to modernize a legacy HRIS by integrating it with Microsoft Azure AD for SSO.
T: I was responsible for ensuring smooth integration without breaking existing workflows.
A: I used Terraform to provision Azure components and built middleware in Python to sync identity data. We rolled out changes incrementally using feature flags and extensive testing in staging.
R: We completed integration with zero downtime and enhanced user experience by reducing login issues by 90%.
Describe your experience automating data flows using APIs or scripting.
S: During my SOC project on AWS, I needed to automate log collection across services.
T: The goal was to centralize data from AWS CloudTrail, GuardDuty, and Security Hub.
A: I used Python and AWS Lambda to pull data via APIs, format it, and forward it to Elasticsearch and Kibana for visualization. I also configured S3 as a backup store.
R: Reduced manual log collection effort by 100% and improved detection of suspicious activities.
How have you contributed to system resilience and fault tolerance?
S: At Eficens, our backend microservices sometimes failed under load.
T: My task was to improve system resilience.
A: I containerized services using Docker, deployed on AWS Lambda behind API Gateway, and implemented retries with exponential backoff in Java. I also added Splunk alerts for anomalies.
R: System uptime improved from 95% to 99.9%, and recovery time after faults dropped by 60%.
💬 BEHAVIORAL QUESTIONS
Tell me about a time you had to work with non-technical stakeholders.
S: During a security project, I worked with the HR team to improve onboarding access flows.
T: They needed a simplified GUI to manage access roles.
A: I conducted a workshop to gather requirements, then built a web-based GUI in Python Flask to manage IAM roles via AWS SDK.
R: Reduced their dependency on IT by 70% and improved provisioning time from 2 days to 1 hour.
Describe a time when you had to manage competing priorities.
S: At TCS, two major features were due at the same time—one for backend optimization, the other for compliance.
T: I was leading both efforts and had to balance development.
A: I prioritized based on risk and regulatory deadline, communicated timelines with stakeholders, and split the team accordingly.
R: Delivered the compliance module on time and delayed the optimization task by only 3 days with no business impact.
🔍 SITUATIONAL QUESTIONS
If your MVP solution starts failing in production, how would you handle it?
S: Imagine a web tool I built for invoice generation starts failing intermittently.
T: My task is to restore service quickly while diagnosing the root cause.
A: First, I’d check logs via Splunk, use health checks, and roll back if a recent deployment caused the issue. I’d isolate the service and implement a retry mechanism. Simultaneously, I’d open comms with stakeholders and document a postmortem.
R: This reduces panic, helps maintain trust, and prevents recurrence via RCA.
What’s your approach to identifying and clearing technical roadblocks?
S: At Eficens, a Lambda function often hit AWS memory limits.
T: I had to fix scalability issues without increasing costs.
A: I refactored logic to batch-process data, reduced cold start times, and used AWS X-Ray to trace bottlenecks.
R: Reduced function runtime by 40% and cost by 25%, unblocking scale-up efforts.
🛠️ TECH STACK QUESTIONS
How have you used PostgreSQL, Elasticsearch, or Snowflake in your past roles?
PostgreSQL: Used it in microservices to manage loan data; wrote optimized queries, used indexing and roles for performance and security.
Elasticsearch: Integrated with Suricata and Zeek logs in my AWS SOC project for threat hunting dashboards.
Snowflake: While I haven’t used Snowflake directly, I’ve worked with similar warehousing platforms and am confident in learning it quickly due to my SQL and ETL experience.
How do you ensure changes are backward compatible?
Use feature flags to enable gradual rollout.
Maintain schema versioning in APIs and DB.
Run regression tests and sandbox testing before deployment.
Use blue-green deployment or canary rollout strategies for critical services.
