Top 30 Architecture Interview Questions and Answers

30 questions. Pick an answer, or open “Show the answer” to read it.

Answers are graded in your browser. Nothing is saved, and no XP or streak is earned here. The app keeps score.

1. Question 1 of 30

   What is the primary function of a Database Management System (DBMS)?

   • aTo develop the user interface and business logic for applications.
   • bTo be the physical collection of all stored data and related files.
   • cTo manage and control how applications interact with and access stored data.
   • dTo provide a structured query language (SQL) for data manipulation.
   Show the answer

   Answer: c · To manage and control how applications interact with and access stored data.

   The card states the DBMS acts as a "single, controlled gateway" and a "librarian" to manage access and operations on data for applications. Option B describes the database itself, which the card identifies as a common misconception, incorrect confusion point.

   Read the full bite: DBMS: The Software That Runs Your Database

2. Question 2 of 30

   Which of the following best describes the primary advantage of adopting a Single Source of Truth (SSOT) for UI components?

   • aIt ensures a consistent user experience and brand identity across all products.
   • bIt guarantees that all UI changes are automatically deployed to every application.
   • cIt significantly reduces the initial setup time for new projects.
   • dIt allows individual teams to customize UI components without affecting others.
   Show the answer

   Answer: a · It ensures a consistent user experience and brand identity across all products.

   The card emphasizes that SSOT solves "the chaos of inconsistent UI" and is critical for maintaining a "consistent look and feel" across multiple applications and teams. While SSOT streamlines development, its core purpose is universal consistency, not individual customization or automated deployment.

   Read the full bite: Single Source of Truth (SSOT): One Place for All UI

3. Question 3 of 30

   What is the primary architectural approach Atlassian's ML Studio uses to balance development velocity with enterprise-level governance?

   • aEnforcing a strict, manual review process for all data access requests and model changes.
   • bAllowing individual teams to select their preferred ML tools and infrastructure for maximum flexibility.
   • cStandardizing the ML lifecycle with reusable modules, centralized orchestration, and automated column-level data governance.
   • dPrioritizing rapid prototyping over compliance to accelerate new feature development.
   Show the answer

   Answer: c · Standardizing the ML lifecycle with reusable modules, centralized orchestration, and automated column-level data governance.

   The card states that the platform "standardizes the ML lifecycle through three core principles: modularity, centralized orchestration, and embedded governance," which includes "automated, column-level data access controls." Option B is incorrect because the platform aims to replace "fragmented tooling" with a "unified platform" and "centralized orchestration," not full team autonomy.

   Read the full bite: Atlassian Details its ML Studio Platform Design

4. Question 4 of 30

   During a bank transfer, if the debit succeeds but the credit fails, which ACID property ensures the money does not vanish?

   • aDurability, because committed changes must survive any system crash
   • bAtomicity, because the entire transaction must complete or roll back entirely
   • cIsolation, because concurrent transactions must not interfere with each other
   • dConsistency, because the database must enforce valid account balances after every update
   Show the answer

   Answer: b · Atomicity, because the entire transaction must complete or roll back entirely

   Atomicity guarantees that a transaction either fully completes or fully rolls back, preventing partial debits without matching credits. Consistency is a tempting distractor because it concerns valid state invariants, not the undo of incomplete operations.

   Read the full bite: Explain ACID properties and why they matter for banking or e-commerce

5. Question 5 of 30

   For which type of task would Node.js's event-driven model, relying on EventEmitter, typically be least effective?

   • aPerforming intensive image manipulation or video encoding.
   • bHandling asynchronous file system operations like reading or writing.
   • cManaging thousands of simultaneous client connections to a web server.
   • dProcessing large data streams from a database query.
   Show the answer

   Answer: a · Performing intensive image manipulation or video encoding.

   The event-driven model is least effective for CPU-bound tasks like image manipulation because they block the single-threaded event loop. It excels at I/O-bound tasks such as network requests, database queries, and file operations.

   Read the full bite: Node.js Events and the EventEmitter

6. Question 6 of 30

   What is the primary reason to avoid using "natural keys" like email addresses as primary keys?

   • aTheir values can change over time, complicating data management and relationships.
   • bThey are typically longer strings, which can slow down database queries.
   • cThey often contain sensitive personal information, posing security risks.
   • dThey cannot be automatically generated by the database system.
   Show the answer

   Answer: a · Their values can change over time, complicating data management and relationships.

   The card states that natural keys "can change, which creates a maintenance nightmare" and requires updating the value in every referencing table. This directly points to complications in data management and maintaining relationships. While other options might have some truth in different contexts, the card emphasizes changeability as the core issue.

   Read the full bite: Primary Keys: The Unique ID for Every Row

7. Question 7 of 30

   A CEO announces a strategic pivot and asks engineering to begin replanning immediately. What is the most appropriate first response?

   • aPropose a full rewrite to avoid carrying legacy constraints into the new strategy
   • bDefine kill criteria and quarterly checkpoints to enable reversal if the market shift fails
   • cAudit current architecture for coupling and technical debt that could block the pivot
   • dMeet with product and executives to define success metrics, timelines, and constraints before evaluating systems
   Show the answer

   Answer: d · Meet with product and executives to define success metrics, timelines, and constraints before evaluating systems

   The card states that the first step is to clarify new business outcomes with product and executives before touching any code. Auditing architecture is step two and requires that context first, while defining kill criteria is the final step; proposing a rewrite immediately is the biggest red flag.

   Read the full bite: How do you assess architecture impact during a corporate pivot?

8. Question 8 of 30

   What primary challenge did the relational model address compared to earlier database systems?

   • aDifficulty in representing data with complex, hierarchical relationships
   • bInability to handle concurrent access from multiple applications
   • cThe need for complex procedural code to navigate specific data storage paths
   • dLack of support for storing unstructured or semi-structured data types
   Show the answer

   Answer: c · The need for complex procedural code to navigate specific data storage paths

   The card states that before the relational model, "Retrieving data required writing complex procedural code to navigate specific data paths," which was inflexible. The relational model aimed to separate the logical data structure from its physical implementation. Option A is a distractor because while hierarchical models were predecessors, the core problem was the procedural navigation within them, not just the representation of hierarchies.

   Read the full bite: The Relational Model: Data as Simple Tables

9. Question 9 of 30

   Why does Angular treat dependency injection as a runtime architectural layer rather than relying on ES module imports like Vue?

   • aIt automatically removes unused services from the bundle regardless of whether they are listed in module providers.
   • bIt avoids constructor boilerplate by letting components request dependencies without declaring them in constructors.
   • cIt enables runtime resolution of service lifetimes and implementation swapping without changing consumer code or direct file coupling.
   • dIt provides a hierarchical state container identical to Vue's provide and inject API.
   Show the answer

   Answer: c · It enables runtime resolution of service lifetimes and implementation swapping without changing consumer code or direct file coupling.

   Angular's DI is a runtime inversion-of-control system that supports scoped lifetimes, tree-shakable providers, and test-time substitution without modifying consumer code. The boilerplate distractor is wrong because DI still requires constructor declarations; its architectural value lies in runtime indirection and hierarchical resolution, not syntax savings.

   Read the full bite: Why is DI central to Angular architecture versus Vue's module system?

10. Question 10 of 30

    Why are positional encodings necessary before the first self-attention layer in a standard Transformer?

    • aBecause self-attention computes unordered pairwise dot products, leaving attention scores unchanged when tokens are permuted
    • bBecause residual connections can only propagate positional information if encodings are concatenated to embeddings
    • cBecause multi-head attention uses position-dependent scaling factors that require bounded inputs
    • dBecause the feed-forward network otherwise infers sequence order implicitly from training data
    Show the answer

    Answer: a · Because self-attention computes unordered pairwise dot products, leaving attention scores unchanged when tokens are permuted

    Self-attention is permutation-invariant: swapping tokens does not change their dot-product attention scores, so the model receives no sequence-order signal without positional encodings. Distractor A represents the common misconception that feed-forward layers can learn implicit order, but they process each position independently and cannot recover shuffled order.

    Read the full bite: Explain positional encodings in Transformers and their necessity

11. Question 11 of 30

    When scoping an MLOps platform for a mid-sized company with 5-15 engineers, which approach best demonstrates mature build-vs-buy reasoning?

    • aPrioritize data governance, feature store, model registry, CI/CD/CT, and monitoring before the serving layer; buy commodity tools like orchestration and monitoring while investing engineering effort only in proprietary feature engineering and model architectures.
    • bBuild a custom Kubernetes orchestrator and real-time GPU serving cluster to minimize latency, deferring data lineage tooling until the team grows.
    • cBuild the feature store and model registry in-house to avoid vendor lock-in, using open-source monitoring with minimal configuration to keep costs low.
    • dPurchase managed feature engineering pipelines and model training platforms, but build custom experiment tracking and observability stacks because production visibility is a core differentiator.
    Show the answer

    Answer: a · Prioritize data governance, feature store, model registry, CI/CD/CT, and monitoring before the serving layer; buy commodity tools like orchestration and monitoring while investing engineering effort only in proprietary feature engineering and model architectures.

    This option correctly sequences foundational components before serving and applies the buy-commodity, build-differentiator rule. Option C is tempting because avoiding vendor lock-in feels engineering-savvy, but maintaining a custom feature store and registry would consume two to three full-time engineers and ignores total cost of ownership.

    Read the full bite: Design an MLOps platform for a mid-sized company: components and build-vs-buy trade-offs

12. Question 12 of 30

    In a bank transfer, which ACID property guarantees that if the transaction fails midway, both accounts revert to their original state?

    • aIsolation
    • bConsistency
    • cDurability
    • dAtomicity
    Show the answer

    Answer: d · Atomicity

    Atomicity ensures that a transaction is treated as an 'all or nothing' operation; if any part fails, the entire transaction is rolled back. Consistency, while ensuring the database moves from one valid state to another, does not specifically handle the rollback of an incomplete operation in the same way Atomicity does.

    Read the full bite: ACID: The Four Guarantees of Database Transactions

13. Question 13 of 30

    What key technical foundation allows a team to implement frequent changes in a disciplined, Agile manner, rather than descending into chaos?

    • aA well-groomed backlog and daily stand-up meetings to track progress.
    • bLoosely coupled architecture, high test coverage, and a mature CI/CD pipeline.
    • cA change control board that requires detailed upfront specifications for all modifications.
    • dA culture of working overtime to immediately implement any new stakeholder request.
    Show the answer

    Answer: b · Loosely coupled architecture, high test coverage, and a mature CI/CD pipeline.

    Disciplined agility relies on a technical foundation that makes change safe and cheap, such as loose coupling and CI/CD. A common misconception is that Agile is only about process; while backlog grooming is important, it doesn't technically enable safe, rapid change.

    Read the full bite: Agile Change vs. Chaos: Technical Enablers

14. Question 14 of 30

    Agile's principle of "welcoming change" is distinguished from chaotic development primarily by its reliance on what?

    • aEmpowering teams to make independent decisions without external oversight.
    • bFrequent stakeholder meetings and rapid iteration cycles.
    • cMinimizing detailed documentation and upfront design.
    • dA robust framework of engineering discipline, including automated testing and modular design.
    Show the answer

    Answer: d · A robust framework of engineering discipline, including automated testing and modular design.

    The card explicitly states that Agile's flexibility is a direct result of rigorous engineering discipline, highlighting practices like comprehensive automated testing, CI/CD, and modular architectural design. While frequent meetings and rapid iterations (Option B) are Agile practices, without this underlying technical discipline, they can lead to chaos rather than effective change management.

    Read the full bite: How does Agile's 'welcoming change' differ from chaos?

15. Question 15 of 30

    Which statement best describes how disciplined Agile teams keep changing requirements from creating chaos?

    • aThey rely on short sprints and frequent standups to detect changes early.
    • bThey combine collaborative feedback loops with engineering practices that make change cheap and safe.
    • cThey minimize upfront planning so the team can pivot immediately without architectural constraints.
    • dThey accept only changes that fit within the original scope to protect the schedule.
    Show the answer

    Answer: b · They combine collaborative feedback loops with engineering practices that make change cheap and safe.

    The card emphasizes that Agile welcomes change through technical guardrails like automated testing, continuous integration, and evolutionary architecture, not just process ceremonies. Option A is tempting because it names real Agile practices, but the card explicitly warns that standups and sprints alone do not make change safe.

    Read the full bite: How does Agile welcome changing requirements without chaos?

16. Question 16 of 30

    When a service is provided within a specific component's providers array, what is the primary outcome?

    • aAngular will always throw a NullInjectorError if the service is not also provided at the root.
    • bThe service becomes a global singleton, accessible throughout the entire application.
    • cThe service can only be injected by the component itself, and not by any of its child components.
    • dA new instance of the service is created for that component and its descendants, isolated to that part of the UI.
    Show the answer

    Answer: d · A new instance of the service is created for that component and its descendants, isolated to that part of the UI.

    Providing a service at the component level creates a new instance of that service, scoped to that component and its children, allowing for isolated state within a UI branch. This differs from `providedIn: 'root'`, which creates an application-wide singleton.

    Read the full bite: Angular's Hierarchical Dependency Injection

17. Question 17 of 30

    What is the main objective of applying database normalization in a relational database?

    • aTo ensure data integrity and minimize redundancy by structuring data into smaller, related tables.
    • bTo facilitate the storage and retrieval of unstructured data types, such as documents or media files.
    • cTo consolidate all related data into a single, comprehensive table for easier access.
    • dTo optimize complex analytical queries by reducing the need for JOIN operations.
    Show the answer

    Answer: a · To ensure data integrity and minimize redundancy by structuring data into smaller, related tables.

    Normalization's primary goal is to eliminate data redundancy and prevent modification anomalies (update, insertion, deletion), thereby ensuring data integrity. Option D describes a benefit of denormalization, often used in analytical systems, which is the opposite of normalization's typical effect on JOINs.

    Read the full bite: Database Normalization: Tidy Tables, Less Redundancy

18. Question 18 of 30

    Which fundamental design aspect of the React Native Bridge is primarily responsible for its performance limitations and visible UI jank?

    • aThe necessity of compiling JavaScript code to native machine instructions.
    • bIts asynchronous message passing mechanism between the JavaScript and native threads.
    • cThe overhead introduced by serializing and deserializing all messages into JSON format.
    • dThe requirement for all UI components to be rendered directly by the JavaScript thread.
    Show the answer

    Answer: b · Its asynchronous message passing mechanism between the JavaScript and native threads.

    The card explicitly states that the Bridge's asynchronous nature and the inherent delays it causes are the source of its main limitations, leading to UI jank and preventing synchronous UI layout access. While JSON serialization adds overhead, the core problem causing delays and visual 'jumps' is the asynchronous communication.

    Read the full bite: The React Native Bridge: Why It's Being Replaced

19. Question 19 of 30

    To model a state machine where each state might carry unique, type-specific data (e.g., a success state holding a user object, an error state holding a message), which Kotlin construct is most appropriate?

    • aAn enum class
    • bA sealed class
    • cAn interface
    • dAn abstract class with a private constructor
    Show the answer

    Answer: b · A sealed class

    Sealed classes are specifically designed for representing restricted hierarchies where subclasses can be distinct types (like data classes or objects) and hold different associated data. Enum classes cannot hold different types of data for each constant, making them unsuitable for states with unique payloads.

    Read the full bite: When to use a sealed class instead of an enum?

20. Question 20 of 30

    To model a network request's state—`Loading`, `Success` (with data), or `Error` (with a message)—which Kotlin construct is most suitable?

    • aAn enum, because it represents a fixed set of constants, which is what these states are.
    • bA sealed interface, because it provides maximum flexibility if the states need to inherit from other classes.
    • cAn enum with nullable properties for both data and an error message to accommodate all states.
    • dA sealed class, because each of its subclasses can carry different data types specific to that state.
    Show the answer

    Answer: d · A sealed class, because each of its subclasses can carry different data types specific to that state.

    A sealed class is ideal because its subclasses can represent states with different associated data (e.g., data for Success, a message for Error) in a type-safe way. An enum is a set of constants of the same type and cannot elegantly model states with different data requirements.

    Read the full bite: When would you use a sealed class instead of an enum?

21. Question 21 of 30

    By default, what action does a relational database take when a user attempts to delete a record from a "parent" table that is referenced by a foreign key in a "child" table?

    • aIt converts the foreign key in the child table into a primary key to preserve the relationship.
    • bIt automatically sets the foreign key column in the child table to NULL for all referencing records.
    • cIt blocks the deletion of the parent record to maintain data integrity.
    • dIt automatically deletes all corresponding referencing records in the child table.
    Show the answer

    Answer: c · It blocks the deletion of the parent record to maintain data integrity.

    The card explicitly states that if you try to delete a user referenced by posts, "the database will, by default, block the deletion to avoid creating orphaned posts." This is the standard behavior to enforce referential integrity. Option B describes an 'ON DELETE SET NULL' action, which is a configurable behavior, not the default.

    Read the full bite: Foreign Keys: The Glue of Relational Databases

22. Question 22 of 30

    Which cloud strategy combination best addresses predictable monthly user growth and unpredictable, short-duration traffic spikes?

    • aRely solely on elasticity to handle both the monthly growth and celebrity-induced spikes.
    • bImplement scalability for the monthly growth and elasticity for the celebrity-induced spikes.
    • cUse scalability exclusively, pre-provisioning resources for both anticipated growth and potential spikes.
    • dEmploy a fixed resource allocation, manually adjusting capacity only when performance degrades.
    Show the answer

    Answer: b · Implement scalability for the monthly growth and elasticity for the celebrity-induced spikes.

    Scalability is ideal for planned, long-term growth, while elasticity efficiently handles unpredictable, short-term demand fluctuations by automatically adjusting resources. Relying only on elasticity for steady growth is inefficient, as it would constantly make small adjustments instead of operating from a stable baseline.

    Read the full bite: Cloud Scalability vs. Elasticity: Planned Growth vs. Real-Time Reaction

23. Question 23 of 30

    Why should the service layer in a layered Express API avoid referencing the request and response objects?

    • aServices run in a different thread from controllers
    • bExpress forbids accessing them outside route handlers
    • cPassing them is slower than passing plain values
    • dIt keeps business logic framework-agnostic and unit-testable without HTTP
    Show the answer

    Answer: d · It keeps business logic framework-agnostic and unit-testable without HTTP

    Keeping req and res out of services makes the business logic reusable and testable without spinning up HTTP. Express does not forbid it and services are not on a separate thread.

    Read the full bite: Layered structure for a scalable Express API

24. Question 24 of 30

    What is the main performance trade-off introduced by adding a database index?

    • aIt requires manual updates to maintain data consistency.
    • bIt increases the complexity of database schema design.
    • cIt makes queries on non-indexed columns significantly slower.
    • dIt speeds up data retrieval but slows down data modification operations.
    Show the answer

    Answer: d · It speeds up data retrieval but slows down data modification operations.

    The card explicitly states that indexes speed up data retrieval but add overhead to INSERT, UPDATE, and DELETE operations, as the index structure must also be modified. This trade-off between faster reads and slower writes is the primary consideration. Indexes do not inherently complicate schema design, slow down non-indexed queries, or require manual updates for consistency.

    Read the full bite: Database Index: The Phonebook for Your Data

25. Question 25 of 30

    When a B2B SaaS company decides to build its own customer-facing analytics platform, which long-term challenge is most frequently underestimated compared to initial development costs?

    • aThe ongoing engineering effort for maintenance, scaling, security, and new feature development, diverting resources from core product.
    • bThe risk of the solution becoming technologically obsolete due to rapid changes in analytics technology.
    • cThe initial developer salaries and infrastructure setup costs.
    • dThe difficulty in achieving feature parity with leading third-party analytics platforms.
    Show the answer

    Answer: a · The ongoing engineering effort for maintenance, scaling, security, and new feature development, diverting resources from core product.

    The card highlights that the 'true cost' of building is the perpetual maintenance burden, including scaling, security, and ongoing feature development, which diverts engineering resources from core product work. Focusing only on initial costs like developer salaries is a common misconception that misses these significant long-term burdens.

    Read the full bite: Build vs. Buy: Third-Party vs. In-House Analytics

26. Question 26 of 30

    Which principle is central to High Availability's ability to maintain continuous service during component failures?

    • aImplementing redundant components with automatic failover mechanisms to handle detected outages.
    • bRelying on human operators to quickly identify and manually switch to backup systems when issues arise.
    • cEliminating all potential single points of failure through exhaustive pre-deployment testing.
    • dDistributing system workloads evenly across a large cluster of independent, non-redundant servers.
    Show the answer

    Answer: a · Implementing redundant components with automatic failover mechanisms to handle detected outages.

    The core of High Availability involves having redundant components and automatically rerouting traffic to a healthy standby when a primary component fails. Distractor B is incorrect because HA specifically emphasizes automatic failover, not manual intervention.

    Read the full bite: High Availability: Designing Systems That Don't Go Down

27. Question 27 of 30

    What is the primary distinction between a standard database view and a base table?

    • aA view executes a predefined query against base tables each time it is accessed, without storing its own data.
    • bA view can only be used to restrict access to specific rows, while a base table allows full access.
    • cA view physically stores its own copy of the data, unlike a base table.
    • dA view automatically optimizes complex joins, making it inherently faster than querying base tables directly.
    Show the answer

    Answer: a · A view executes a predefined query against base tables each time it is accessed, without storing its own data.

    A standard view stores only its query definition and executes it against the underlying base tables each time it's accessed, as explained in the 'How It Works' section. Option C describes a common misconception; views do not store their own data, unlike materialized views.

    Read the full bite: Database Views: A Saved Query That Acts Like a Table

28. Question 28 of 30

    The cloud's Region/Availability Zone (AZ) model is primarily designed to address which of the following challenges?

    • aThe requirement for applications to automatically scale up and down based on demand.
    • bThe inherent risk of a single data center becoming a critical point of failure.
    • cThe need to distribute application load evenly across a global user base.
    • dThe complexity of managing virtual machines and containers in a distributed environment.
    Show the answer

    Answer: b · The inherent risk of a single data center becoming a critical point of failure.

    The card states that "A single data center is a massive single point of failure" and the Region/AZ model was created to allow applications to "survive the loss of an entire data center or more." This directly addresses the risk of single data center failures. While other options are general cloud benefits, they are not the primary design goal of the Region/AZ model's failure isolation.

    Read the full bite: Cloud Regions and AZs: Infrastructure Built for Failure

29. Question 29 of 30

    A startup that has not yet found product-market fit targets a winner-take-all market and asks whether to use microservices. Which response best aligns architecture with its business risk profile?

    • aStart with microservices to avoid a costly rewrite later, since technical debt in a monolith compounds exponentially.
    • bStart with microservices if you have strong DevOps practices, otherwise default to a monolith to minimize cloud infrastructure spending.
    • cStart with a monolith to validate hypotheses cheaply and preserve speed, extracting services only after bounded contexts hit concrete scaling or team-size pain.
    • dBuild a monolith but immediately extract every bounded context into a separate service to prevent the tight coupling that doomed early Amazon.
    Show the answer

    Answer: c · Start with a monolith to validate hypotheses cheaply and preserve speed, extracting services only after bounded contexts hit concrete scaling or team-size pain.

    The card frames a monolith as cheap optionality under uncertainty and warns that microservices before product-market fit impose a scaling tax; in winner-take-all markets, speed is existential. Option A is tempting because engineers often fear rewriting, but it treats microservices as a universal best practice and ignores the operational overhead and business risk profile described in the card.

    Read the full bite: How do you frame monolith vs microservices trade-offs under market uncertainty?

30. Question 30 of 30

    Why is a W3C-compliant JSON file preferred over Figma or CSS as the single source of truth in a cross-platform token pipeline?

    • aFigma and CSS are not version-controlled artifacts and CSS is platform-specific
    • bCSS variables lack support for color values on mobile platforms
    • cStyle Dictionary requires JSON input to generate Swift enums for iOS
    • dJSON files can be edited directly by designers without engineering involvement
    Show the answer

    Answer: a · Figma and CSS are not version-controlled artifacts and CSS is platform-specific

    The correct answer captures the governance requirement that the source of truth must be version-controlled and platform-agnostic, which Figma and CSS cannot guarantee. Distractor D is tempting because Style Dictionary is part of the recommended pipeline, but JSON is chosen for immutability and cross-platform neutrality, not merely because the build tool accepts it.

    Read the full bite: Design token pipeline serving Web, iOS, and Android