Blackrock Online Assessment MCQS (SWE + Aladdin Data Role) 8 August 2026 - Manipal University - BlackRock Interview Question

Blackrock Online Assessment MCQS (SWE + Aladdin Data role)

1. Analytical

Question: Starting at (0,0), you want to reach (4,4) by moving only Right or Up one unit at a time. Coordinates (1,1) and (3,3) contain landmines and cannot be used. How many completely safe paths exist?

2. Aptitude - Diagnostic

Question: A diagnostic test detects a condition correctly 90% of the time when the condition is present, and correctly returns negative 80% of the time when the condition is absent. If 10% of people have the condition, what is the probability that a person actually has the condition given that the test result is positive?

3. Aptitude - Defective chip

Question: A batch has 4 defective chips and 8 good chips. Two chips are picked sequentially without replacement. What is the probability that both are defective?

1/11
1/9
2/11
3/22

4. Problem Solving (12-Queens)

Question: In the 12-Queens problem, after placing the first queen somewhere in row 1, how many possible positions are there for placing the next queen in row 2 before checking diagonal conflicts?

12
24
144
1024

5. Logical Reasoning (Snacks)

Question: People P, Q, R and S each like one different snack among Sandwich, Noodles, Pasta and Salad. Clues: P does not eat Salad. Q eats Noodles. R does not eat Pasta. S eats Salad. Who eats Sandwich?

6. Aptitude (Spider)

Question: A spider starts at one vertex (corner) of a wireframe cube. Every minute, the spider randomly chooses one of the three adjacent edges and crawls to the next vertex. What is the expected number of moves needed to reach the vertex diagonally opposite the starting point?

7. Aptitude (Pill Bottles)

Question: Ten bottles contain pills. One bottle contains pills that are 2 grams heavier. You may use a digital scale exactly once. What is the minimum total number of pills you need to weigh to identify the heavier bottle?

8. Logical Reasoning (Bridge Crossing)

Question: Four people take 1, 4, 7 and 11 minutes respectively to cross a bridge. A torch is required to cross the bridge, at most two people can cross at a time, and a pair moves at the slower speed compared to a single person's speed. What is the minimum total time needed for all to cross?

9. Problem Solving (Horses)

Question: There are 25 horses and only 5 can race at a time. You only get the relative ranking within each race, with no timings. What is the minimum number of races required to identify the top 4 horses?

10. Critical Reasoning - Paradox

Question: In a company, the CTO says: "There exists at least one engineer who, if they know the system is secure, then the system is secure." This seems trivially true. However: No engineer actually knows whether the system is secure. Why does the statement still hold?

The statement relies on an implication that is vacuously true
Engineers are assumed to be omniscient
The CTO's statement is logically flawed
Security cannot be proven

11. Aptitude (Permutations)

Question: How many permutations of the letters W, X, Y, Z, A, B satisfy all of the following conditions? W is not in position 1, XXX and YYY are not adjacent, and ZZZ appears before AAA?

12. Critical Reasoning - Paradox (Surprise Test)

Question: A professor announces: "There will be exactly one surprise test next week (Monday-Friday), and you will not be able to predict the day beforehand." Students reason it cannot be Friday, then Thursday, and eliminate all days. Conclusion: No test is possible. However, the test is on Wednesday, and students are surprised. What explains this paradox?

Students made a logical fallacy in backward induction
The paradox proves prediction itself is impossible
Surprise is subjective, so logic does not apply
The reasoning assumes perfect knowledge that students do not have

Software Engineering & Spring

13. Spring Data Repositories

Question: Given a users table and a Spring Data JPA Repository interface extending JpaRepository<User, Integer>. What is the result of using a custom @Query("update User u set u.age = ?2 where u.id = ?1") void setAge(int id, int age); in the Repository and calling it from a Service?

setUserAge method runs fine and sets user matching id age to the specified age value.
setUserAge throws InvalidDataAccessApiUsageException when executed
setUserAge method runs fine... if @Transactional(readOnly = false) is inserted
setUserAge method runs fine... if @Modifying and @Transactional are inserted

14. Data Mapping

Question: Design a mapping to load every 200th record to the target table, e.g., if the source has 600 records, load only the 200th, 400th, and 600th. In the implementation steps, what should replace Step 3?

Make a new output field of type integer Flag with the value MOD(V_count,200). a.V_count=IIF(V_count=200,1,V_count+1) b.Flag=V_count
Make a new output field of type integer Flag with the value MOD(V_count,200). a.V_count=V_count+1 b.Flag=MOD(V_count,200)
Make a new output field of type integer Flag with the value MOD(V_count,200). a.V_count=V_count-1 b.Flag=MOD(200,V_count)
Make a new output field of type integer Flag with the value MOD(V_count, 200). a.V_count+1=V_count b.MOD(V_count,200)=Flag

Databases & SQL Concepts

15. Normal Form

Question: Given the following functional dependencies, find the highest normal form. Assume that the given relation is in 1NF already. R(A,B,C,D,E,F). Functional Dependencies: BC->E, BE->F, AF->DE.

1NF
2NF
3NF
BCNF

16. Aggregation Optimization (Databricks)

Question: There is a need to optimize a Databricks Delta Lake table sensor data for frequent analytical aggregations. Which implementation applies advanced optimizations (ZORDER, Caching, Adaptive Query Execution)?

... sensor_data.write.option("zorder_cols", "sensor_id")... sensor_data_zordered.cache() spark.conf.set("spark.sql.adaptive.enabled", "true") ...
(Other code variations omitted for brevity but conceptually represented as incorrect choices)

17. Data Concepts 16

Question: Which query returns departments with more than 5 employees?

SELECT department_id FROM employees WHERE COUNT(*) > 5 GROUP BY department_id;
SELECT department_id FROM employees GROUP BY department_id HAVING COUNT(*) > 5;
SELECT department_id FROM employees HAVING COUNT(*) > 5;
SELECT department_id FROM employees WHERE employees > 5;

18. Data Concepts 18

Question: What does this query return? SELECT department_id, AVG(salary) FROM employees GROUP BY department_id;

Average salary across all employees
Average salary per department
Highest salary per department
Number of employees per department

19. Data Concepts 17

Question: What is wrong with this query? SELECT department_id, employee_name, COUNT(*) FROM employees GROUP BY department_id;

COUNT(*) cannot be used with GROUP BY
employee_name is selected but not grouped or aggregated
department_id cannot be grouped
Query must include ORDER BY

20. Data Concepts 19

Question: Which aggregate function ignores NULL values?

AVG()
SUM()
COUNT(column)
All of the above

21. Data Concepts 20

Question: Which query finds duplicate email addresses?

SELECT email FROM users WHERE COUNT(email) > 1;
SELECT email, COUNT() FROM users GROUP BY email HAVING COUNT() > 1;
SELECT DISTINCT email FROM users WHERE email > 1;
SELECT email FROM users GROUP BY COUNT(email);

22. Data Concepts 66

Question: Which query helps detect duplicate primary key values?

SELECT id, COUNT() FROM table_name GROUP BY id HAVING COUNT() > 1;
SELECT id FROM table_name WHERE id IS NULL;
SELECT DISTINCT id FROM table_name;
SELECT COUNT(id) FROM table_name;

23. Data Concepts 67

Question: Which query detects unexpected NULLs in a required column?

SELECT * FROM table_name WHERE required_column IS NULL;
SELECT * FROM table_name WHERE required_column = NULL;
SELECT * FROM table_name WHERE required_column LIKE NULL;
SELECT * FROM table_name GROUP BY required_column;

24. Data Concepts 68

Question: Which technique helps validate that a child table does not reference missing parent records?

Anti-join child table to parent table
Sort both tables by name only
Count all rows in parent table
Use SELECT DISTINCT * only

25. Serializability

Question: Consider the transactions (T1, T2) and schedules (S1, S2, S3). Which statement is true?

S1 is conflict serializable
S1 is conflict equivalent to S2
S1 is conflict equivalent to S3
S3 is conflict serializable

26. Candidate Key

Question: Given the functional dependencies: 1. B->CD, 2. ACD->BE, 3. BD->EF, 4. D->G, 5. H->G for R(A,B,C,D,E,F,G,H). Find the candidate key(s).

ABEFGH
ABH
AB
AH

27. PL/SQL: FORALL

Question: 10000 records need transferring using FORALL. One bad record causes failure. Solution?

FORALL cannot be used for insertion
Use SAVE EXCEPTIONS
Select the bad data first and remove it
Use CURSOR for row-by-row processing

28. Database Operation (JDBC)

Question: Which code block will update the emp_data column of the EMPLOYEE table using JDBC?

(Code block with ResultSet.CONCUR_UPDATABLE)
(Code block with ResultSet.FETCH_FORWARD)
(Code block with ResultSet.TYPE_SCROLL_SENSITIVE)
None of the above

29. Practical database partitioning

Question: What is a good way to partition a database of users having just first name, last name, and unique integer?

Partition users by first letter of their names
Partition users by hash of the first letter
Partition users by their identifiers modulo m
Partition users by a hash of their identifiers modulo m

30. Errors in a Block (PL/SQL)

Question: In PL/SQL code...

If the current block is not having the exception handler the enclosing block will be searched for one
If an error occurs in the DECLARE section of a block, immediately the enclosing block is searched
If an error occurs in the exception section of a block immediately the enclosing block is searched
none of these

31. Database normalization

Question: In general, which of the following are true about database normalization?

Enforces better data integrity at the cost of greater storage space
Enforces better data integrity at the cost of reduced performance in data retrieval
Reduces used storage space
Performing updates to the normalized data is significantly easier and has better performance

32. Database Selection

Question: A ride-hailing company with 50,000 drivers handles 2 million trips/day... needs to respond with current locations in < 50ms. Which database?

LevelDB
CouchDB
An external SQL database
An external key-value store like Cassandra

33. Identify Logical Database Structures

Question: Which of the following are logical structures in a database?

Redo logs
Data files
Data blocks
Control files

Cloud Architecture & Security (AWS/General Cloud)

34. S2 Q14 - Load balancer

Question: What does a load balancer do?

Encrypts all databases
Stores files
Distributes traffic across backend targets
Generates AI embeddings

35. S2 Q23 - Egress risk

Question: Why should data egress be considered in cloud architecture?

It can introduce cost, latency, and security considerations
It is always free and instant
It only affects UI design
It prevents encryption

36. S2 Q24 - Secrets in logs

Question: What is the best reason to avoid logging tokens and secrets?

Logs are never stored
Logs may be widely accessible or retained, causing credential leakage
Secrets make logs too short
Logging secrets improves debugging safely

37. S2 Q25 - Cloud key rotation

Question: What is secret or key rotation?

Changing cloud regions
Compressing storage keys
Making keys public
Periodically replacing credentials to reduce exposure

38. S2 Q28 - Public bucket risk

Question: What is the biggest risk of accidentally public object storage containing internal files?

Files become compressed
Unauthorized data exposure
AI models stop working
The bucket becomes read-only

39. S2 Q29 - WAF

Question: What is a web application firewall primarily used for?

Replacing unit tests
Storing large files
Running GPU workloads
Filtering and blocking malicious web traffic patterns

40. S2 Q20 - Network segmentation

Question: What is network segmentation intended to reduce?

Use of private IP addresses
Blast radius by limiting communication between systems
The number of developers
Data durability

41. S2 Q13 - Security group/firewall

Question: What is the purpose of a cloud firewall or security group?

To increase memory
To classify prompts
To create object storage
To control allowed inbound and outbound traffic

42. S2 Q12 - Private subnet

Question: Why place databases in private subnets?

To make them easier to access publicly
To reduce direct exposure to the internet
To disable backups
To avoid authentication

AI & Machine Learning Integrations

43. S2 Q19 - Model output validation

Question: Why should structured AI outputs be validated programmatically before use?

The model may produce malformed, missing, or semantically invalid fields
Validation makes output creative
Validation trains the model
Validation removes all need for tests

44. S2 Q17 - Data leakage

Question: Which scenario is the clearest example of AI data leakage?

AI summarizes public documentation
AI refuses an unsafe request
AI asks for clarification
A chatbot reveals confidential client data to an unauthorized user

45. S2 Q16 - AI access control

Question: Why must RAG retrieval enforce document-level permissions?

To make answers longer
To avoid exposing documents the user is not allowed to see
To improve temperature settings
To skip indexing

46. S2 Q05 - Citation value

Question: Why are citations useful in a RAG system?

They help users verify the source of claims
They make all answers correct
They reduce network latency
They eliminate access control

47. S2 Q04 - Prompt injection basic

Question: What is prompt injection?

Adding examples to improve output
Malicious or unintended input that tries to override instructions
Compressing a prompt
Tokenizing a sentence

48. S2 Q03 - Human review

Question: When is human review especially important for AI output?

Only when output is creative
Never, if the model is large
Only when the answer is short
When decisions affect people, finances, compliance, or production systems

49. S2 Q02 - Bias

Question: What is bias in AI output?

A billing report
A vector database error
A systematic unfair or skewed pattern in responses
A file storage format

50. S2 Q01 - Sensitive data in prompts

Question: Why should employees avoid putting sensitive client or internal data into unmanaged AI tools?

AI cannot read sensitive text
It may be logged, exposed, or processed outside approved controls
Sensitive data always improves accuracy
It prevents hallucination

Design Patterns & Principles

51. DP-H-025 (Template Method)

Question: What is the principal risk of Template Method's hook methods?

Subclasses cannot call the template method
Subclasses can break the algorithm by overriding non-abstract hook methods with incompatible behavior
The template method cannot call private methods
The algorithm structure is visible to clients

52. DP-H-067 (Spring AOP)

Question: How does Spring AOP use the Proxy pattern, and what is the limitation of interface-based proxies?

Spring AOP uses Decorator; no limitations
Spring AOP creates JDK dynamic proxies (for interfaces) or CGLIB proxies (for classes); JDK proxies only intercept interface method calls, bypass advice
Spring uses Chain of Responsibility for AOP
CGLIB proxies only work with abstract classes

53. DP-H-059 (Singleton Testability)

Question: How does an overused Singleton create testability problems, and what is the recommended fix?

Singletons are always thread-unsafe, so tests fail; fix by synchronizing
Singletons introduce global shared state that persists between tests; fix by injecting the dependency via an interface/test double
Singletons cannot be instantiated in test frameworks; fix by using reflection
Singletons prevent mocking because they are final

54. DP-H-033 (SOLID)

Question: A UserService handles persistence, email notification, and authentication. Which SOLID principle is violated and what is the correct fix?

OCP; split into user, email, auth services connected by interfaces
SRP; refactor into separate UserRepository, EmailService, and AuthService classes
DIP; inject dependencies into a single UserService
ISP; create separate interfaces for each responsibility

55. DP-M-054 (Repository Pattern)

Question: The Repository pattern builds primarily on which GoF pattern?

Command
Observer
Facade (abstracting data access behind a collection-like interface)
Factory Method

56. DP-M-045 (Abstract Factory vs Factory Method)

Question: When should you choose Abstract Factory over Factory Method?

When you need a single product created by client
When you need to create families of related products that must be used together
When the factory needs to be a Singleton
When product creation involves many optional steps

57. DP-M-037 (Dependency Inversion)

Question: The Dependency Inversion Principle requires:

All dependencies to be injected via constructors only
High-level and low-level modules to both depend on abstractions
Avoiding all use of concrete classes
Interfaces to be as large as possible

58. DP-M-005 (Creational Patterns)

Question: Both patterns decouple client code from concrete class names. However, one creates a single product type while the other creates an entire coordinated suite of related products. What is the key structural distinction?

Abstract Factory uses inheritance; Factory Method uses composition
Abstract Factory produces families of related objects; Factory Method produces single type
Abstract Factory is structural; Factory Method is creational
Abstract Factory cannot be combined with Singleton

59. DP-M-067 (Decorator)

Question: Java.io.InputStream wrapped by BufferedInputStream is an example of which pattern?

Proxy
Decorator
Adapter
Facade

60. DP-M-033 (Single Responsibility)

Question: Which SOLID principle states that a class should have only one reason to change?

Single Responsibility Principle
Open/Closed Principle
Interface Segregation Principle
Dependency Inversion Principle

61. DP-M-021 (Observer)

Question: The Observer pattern establishes which relationship between subject and observers?

One-to-one
Many-to-many
One-to-many
Many-to-one

62. DP-M-011 (Adapter)

Question: What is the primary purpose of the Adapter pattern?

Add behavior to objects dynamically
Convert an interface into another interface clients expect
Provide a simplified interface to a complex subsystem
Separate abstraction from implementation

63. DP-M-001 (Singleton)

Question: A distributed logging framework requires every application component to write to the same single log file handle. Which design goal must the class managing this handle satisfy above all others?

Create families of related objects
Clone existing objects
Ensure a class has only one instance and provide a global access point
Define an interface for creating objects

64. DP-S-018 (Template Method)

Question: A data import system supports CSV, JSON, and XML. Each format follows the same pipeline: validate, parse, transform, persist. Only specific steps differ per format. What pattern applies?

Strategy
Chain of Responsibility
Template Method
Builder

65. DP-S-014 (Observer)

Question: A stock trading platform must notify multiple systems (dashboard, alert service, audit log) whenever a trade is executed. Systems must subscribe/unsubscribe dynamically. What pattern applies?

Mediator
Observer
Command
Chain of Responsibility

66. DP-S-010 (Composite)

Question: A financial portfolio system must treat individual positions and groups of positions (sub-portfolios) identically for valuation and risk calculations. What pattern enables this?

Flyweight
Composite
Decorator
Proxy

67. DP-S-073 (Facade)

Question: A legacy ERP system has a complex subsystem for generating financial reports. Your new frontend needs a single clean interface to generate any report type without understanding ERP internals. What applies?

Adapter
Proxy
Facade
Bridge

68. DP-S-069 (Template Method)

Question: A web scraper must parse data from multiple websites. Each site has different HTML structure, but the scraping pipeline (fetch, parse, extract, store) is the same. What pattern fits?

Chain of Responsibility
Template Method
Strategy
Builder

69. DP-S-057 (Observer)

Question: A smart home controller must update the app UI, send push notifications, and log to cloud whenever a device state changes. Devices and consumers must remain decoupled. What pattern applies?

Command
Mediator
Observer
Strategy

70. DP-S-049 (Iterator)

Question: A task scheduler supports multiple queue types (priority queue, FIFO, deadline-based). The scheduler must traverse tasks in different orders without task objects knowing which queue they are in. What pattern applies?

Visitor
Iterator
Strategy
Composite

71. DP-S-029 (Adapter)

Question: You are building an IDE plugin system where plugins are developed by third parties. Plugins must conform to your interface but each has its own internal API. How do you integrate them?

Facade
Adapter
Proxy
Flyweight

72. DP-S-009 (Decorator)

Question: You need to add request logging, authentication, and response compression to an existing API client without modifying its class. These features must be toggleable and combinable. What pattern fits?

Proxy
Adapter
Decorator
Bridge

73. DP-S-045 (Strategy)

Question: A streaming platform runs A/B testing of recommendation algorithms. Users in group A get collaborative filtering; group B gets content-based filtering. The algorithm is swappable per request. What applies?

Template Method
Observer
Abstract Factory
Strategy

74. DP-S-021 (Command)

Question: A ride-sharing app must record every driver action (accept ride, start trip, end trip) so dispatchers can replay events for auditing and the system can retry failed operations. What pattern fits?

Observer
Strategy
Command
Template Method

75. DP-S-001 (Factory Method)

Question: Your application must dynamically support multiple payment options (CreditCard, PayPal, UPI) based on the user's input. How do you architect a system that is clean, scalable, and easy to maintain?

Singleton
Factory Method
Builder
Prototype

76. DP-H-009 (Factory Trade-offs)

Question: What is the core trade-off of using Factory Method over direct instantiation with new?

Factory Method is always slower due to the extra method call
Factory Method decouples the client from concrete types at the cost of a parallel class hierarchy
Factory Method prevents unit testing because the factory cannot be mocked
Factory Method forces the product to be a Singleton

77. DP-H-054 (Strategy Selection)

Question: How would you combine Strategy + Factory + Registry to allow runtime algorithm selection?

Factory creates strategies; Registry selects based on config; Strategy executes
Strategies are registered in a map by key; Factory looks up and returns the appropriate Strategy, client depends only on interface
Registry is a Singleton Factory that returns all strategies
Strategy is a special case of Factory that delegates to a Registry

78. DP-H-045 (Observer Issues)

Question: When does Observer become a poor choice for managing object communication?

When there are more than 10 observers
When update dependencies between objects are complex or circular, making event cascades unpredictable
When observers are defined at compile time
When the subject has only one type of event

Blackrock Online Assessment MCQS (SWE + Aladdin Data role)

1. Analytical

2. Aptitude - Diagnostic

3. Aptitude - Defective chip

Question: A batch has 4 defective chips and 8 good chips. Two chips are picked sequentially without replacement. What is the probability that both are defective?