Top 50 Goldman Sachs MIS Analyst Interview Questions

1. How would you calculate a running total using SQL?

A strong answer should explain the use of a window function such as SUM() OVER() with an appropriate PARTITION BY and ORDER BY.
For example, a running total of monthly sales can be calculated while preserving each individual month's row rather than collapsing the data through a conventional GROUP BY.
The research specifically identifies running totals as a recurring SQL scenario and connects the concept to cumulative exposure, liquidity, and other financial metrics.

2. How would you calculate a 30-day rolling average of trading volume?

Use a window function with an appropriate preceding window, such as a 30-row window where the underlying dataset represents consecutive trading observations.
A strong candidate should also clarify whether the requirement means 30 calendar days or 30 trading days. The research highlights this distinction as an important financial-data consideration because weekends and bank holidays do not represent trading days.

3. How would you identify duplicate transactions when a table has no primary key?

One approach is to identify an immutable combination of business attributes, such as:

• Client ID
• Trade date
• Financial product
• Notional amount

Then use ROW_NUMBER() with PARTITION BY those business keys and order by an ingestion timestamp.
Filtering for ROW_NUMBER() = 1 can preserve one canonical record while identifying subsequent duplicates.

4. What is the difference between WHERE and HAVING in SQL?

WHERE filters individual rows before aggregation, while HAVING filters groups after aggregation.
For example, if you need to identify trading desks whose aggregate losses exceed a threshold, HAVING would be appropriate after grouping the transactions by desk.

5. What is the difference between an INNER JOIN and a LEFT JOIN?

An INNER JOIN returns matching records from both tables.
A LEFT JOIN preserves every record from the left table even when there is no matching record on the right.
In a reconciliation scenario, the distinction can be important. If the primary trade ledger must be preserved even when reference data is missing, a LEFT JOIN may be required.

6. How would you find trades that do not have matching counterparty reference data?

A typical approach would be to use a LEFT JOIN from the trade table to the reference-data table and then filter for rows where the reference-side identifier is NULL.
The research connects this type of query to identifying missing Legal Entity Identifiers and investigating front-office/back-office reconciliation issues.

7. How would you handle NULL values during SQL calculations?

Functions such as COALESCE() can replace missing values with an appropriate fallback value.
The important consideration is not simply eliminating NULLs, but understanding what a missing value means in the underlying business process so that the replacement does not distort the analysis.

8. How would you optimize a SQL query operating on millions of financial records?

A strong answer should focus on reducing unnecessary computation and understanding the query's execution behavior.
Relevant considerations include:

• Filtering appropriately
• Avoiding unnecessary columns
• Choosing suitable joins
• Reducing redundant calculations
• Understanding indexing and data structures where applicable
• Preserving required financial-data granularity

The research emphasizes computational efficiency and the consequences of poorly optimized processing at large scale.

9. How would you calculate the latest record for each employee or client?

A common solution is to use ROW_NUMBER() with PARTITION BY the relevant employee or client identifier and ORDER BY the relevant date or timestamp in descending order.
The first row for each partition represents the latest record.
The supplied research specifically identifies ROW_NUMBER() as useful for finding the latest employee salary as well as deduplicating transaction feeds.

10. How would you count contiguous transaction segments that are strictly increasing?

The research describes a transaction-segment problem in which candidates must count contiguous segments of exactly k transactions that are strictly increasing.
Instead of rebuilding every window from scratch, an optimized sliding-window approach can maintain the current consecutive-increase streak. This reduces unnecessary repeated processing and demonstrates algorithmic efficiency.

11. How would you solve the Two Sum problem efficiently?

The basic problem asks for two numbers in an array whose sum equals a target.
A brute-force solution compares every possible pair and has O(n²) time complexity.
A hash-map approach can store previously seen values and their complements, allowing the problem to be solved in O(n) time with O(n) additional space.
The research specifically identifies Two Sum as an example of the type of algorithmic efficiency expected in the technical screening process.

12. What is Big-O notation, and why is it important for an MIS Analyst?

Big-O notation describes how an algorithm's computational requirements grow as the input size increases.
It matters because MIS systems can process very large financial and operational datasets. An algorithm that works on a small sample can become impractical when applied to millions of records.
The supplied research emphasizes both time and space complexity in the Goldman Sachs technical screening process.

13. How would you solve a Minimum Meeting Rooms problem?

First, sort the intervals by their start times.
Then use a min-heap to track the earliest ending active interval. When a new interval starts before the earliest active interval ends, another room is required.
The research connects this pattern to resource allocation, peak server loads, and overlapping trade execution windows.

14. How would you handle a dataset that suddenly becomes ten times larger?

The key is to reconsider the architecture rather than simply optimize the existing implementation.
A candidate could discuss:

• Memory constraints
• Algorithmic complexity
• Distributed processing
• Streaming approaches
• Data partitioning
• Scalable database processing

The research gives a similar CoderPad scenario in which the interviewer changes the dataset scale and asks the candidate to adapt the solution.

15. How would you approach a complex grid optimization problem?

The research describes an n × m grid problem that can be modeled as a bipartite graph.
The important interview skill is recognizing whether a seemingly complicated problem can be decomposed into independent subproblems and represented using an appropriate mathematical structure.
The supplied research specifically discusses bipartite matching and König's theorem in this context.

16. How do you decide whether an algorithm is suitable for production-scale financial data?

Consider:

• Time complexity
• Space complexity
• Input size
• Memory constraints
• Failure conditions
• Data quality
• Scalability
• Maintainability

Correct output alone is not sufficient when an algorithm is intended for large operational or financial datasets.

17. What is the difference between a brute-force solution and an optimized solution?

A brute-force approach generally explores possibilities directly and may require significantly more computation.
An optimized solution identifies patterns that reduce unnecessary work, such as hashing, sorting, sliding windows, dynamic programming, or appropriate data structures.
The research repeatedly emphasizes moving beyond brute-force solutions during Goldman Sachs technical evaluations.

18. How would you explain your algorithm while solving a live coding problem?

Do not silently type the solution.
A strong approach is to explain:

• Your interpretation of the problem
• Assumptions
• Initial approach
• Complexity
• Why you are changing or optimizing the approach
• Edge cases
• Final solution

This matters because the research describes CoderPad interviews as collaborative sessions where interviewers can actively challenge assumptions and introduce new constraints.

19. Which advanced Excel functions are important for an MIS Analyst?

The research highlights:

• VLOOKUP
• XLOOKUP
• INDEX-MATCH
• Dynamic arrays
• Conditional formatting

Candidates may also be expected to audit complex workbooks and identify logic errors.

20. How would you audit a broken Excel workbook?

Start by identifying:

• The intended business logic
• Incorrect formulas
• Broken references
• Circular references
• Incorrect assumptions
• Data-quality problems
• Dependency relationships

Then test the corrected workbook against known outputs before deploying it.

21. What would you do if an Excel model contains a circular reference?

First determine whether the circular reference is intentional or accidental.
If accidental, trace the formula dependencies and redesign the calculation flow.
The goal is not simply to remove the warning but to ensure that the resulting model follows the intended financial logic.

22. How would you automate a monthly reporting process that currently takes three hours?

Start by mapping the entire workflow.
Identify:

• Data sources
• Manual transformations
• Repetitive Excel operations
• Validation steps
• Formatting requirements
• Distribution requirements

Then determine which parts should be automated using VBA, Python, SQL, or a combination of tools.
The research explicitly identifies this as a key automation scenario.

23. When would you choose VBA instead of Python?

VBA can remain useful when the workflow is heavily Excel-based and requires direct interaction with Excel for users who work primarily within spreadsheets.
Python becomes more appropriate when the workflow involves:

• Large datasets
• Complex transformations
• External APIs
• ETL pipelines
• Greater reproducibility
• Processing beyond Excel's practical limits

The research emphasizes this distinction between Excel-native accessibility and Python's scalability and integration capabilities.

24. How would you use Python for financial data analysis?

A strong answer can discuss Python's role in:

• Data cleaning
• Transformation
• Automation
• Large-scale processing
• ETL workflows
• Reproducible reporting

The supplied research specifically highlights Python and the pandas library when discussing scalable reporting automation.

25. Why can Python be preferable to Excel for large datasets?

Python provides code-based, reproducible workflows and is better suited to datasets and transformations that exceed the practical capabilities of a spreadsheet workflow.
Excel remains valuable for interactive, ad-hoc analysis and user accessibility.
The important point is to choose the technology based on the workflow rather than treating one tool as universally superior.

26. Walk me through the three financial statements.

The three primary financial statements are:

• Income Statement
• Balance Sheet
• Cash Flow Statement

A strong MIS candidate should understand how changes in one statement flow into the others because reporting systems and dashboards depend on the relationships between financial data points.
The research specifically identifies this as a common question crossing into MIS and data-oriented interviews.

27. What happens to the three financial statements if depreciation increases by $10?

The research expects the candidate to trace the change across the statements.
The higher depreciation expense reduces income on the Income Statement. Because depreciation is non-cash, it is added back in the operating section of the Cash Flow Statement, with the research also noting the related tax-shield effect. On the Balance Sheet, PP&E decreases and the corresponding effects flow through retained earnings and cash.

28. What is the difference between Ind AS and IFRS?

A candidate should explain that they are financial reporting frameworks with similarities but differences in specific accounting treatments and requirements.
The research identifies differences between Ind AS and IFRS as a potential interview topic and specifically mentions Ind AS 116 in relation to lease capitalization.

29. Why does an MIS Analyst need accounting knowledge?

MIS reporting can feed financial, operational, regulatory, and management decisions.
Without understanding the accounting meaning behind the data, an analyst may build a technically correct report that represents the underlying financial information incorrectly.
The research describes MIS as intersecting with controllership and financial reporting responsibilities.

30. What is a FOBO break?

A break occurs when financial information between systems or functions does not reconcile.
The research uses the example of Trader P&L from the Front Office differing from Product Control P&L in the Back Office.
The analyst must investigate the underlying trade population, booking information, valuation inputs, and other potential causes rather than simply changing one number to make the systems agree.

31. How would you investigate a difference between Front Office and Back Office P&L?

A structured approach is:

1. Define the break.
2. Identify the affected amount, product, desk, and currency.
3. Reconcile the trade population.
4. Check missing bookings, amendments, dates, or cancellations.
5. Compare market-data inputs.
6. Escalate material discrepancies with evidence.

The research emphasizes maintaining the independence of the official ledger during this process.

32. What is Value at Risk (VaR)?

VaR is a statistical measure used to quantify potential portfolio loss over a specified time period and confidence level.
The research also emphasizes an important limitation: VaR does not adequately describe losses from events beyond the selected confidence threshold, meaning tail risk remains a significant consideration.

33. What are the Greeks in financial markets?

The Greeks are measures of sensitivity used primarily in options risk analysis.
The research identifies:

• Delta
• Gamma
• Theta
• Vega
• Rho

For an MIS Analyst, understanding these concepts can be important when building or interpreting risk dashboards and aggregated portfolio metrics.

34. What is the Liquidity Coverage Ratio?

The Liquidity Coverage Ratio, or LCR, is a Basel III liquidity metric.
The research describes it as relating to a bank's ability to maintain sufficient High-Quality Liquid Assets to survive a severe 30-day stress scenario.

35. How would you build a dashboard when a senior stakeholder gives you vague requirements?

First establish the business objective.
Ask:

• What decision should the dashboard support?
• Who will use it?
• Which metrics matter?
• How frequently should the information update?
• What actions should users take based on the results?

The research recommends rapid prototyping through a wireframe, mockup, or basic Excel pivot before investing heavily in a complex dashboard implementation.

36. What is the difference between a star schema and a normalized database model?

A normalized model is generally structured to reduce redundancy and preserve data integrity.
A star schema typically organizes data around a central fact table with related dimension tables and can simplify analytical querying.
The research specifically expects MIS candidates to understand why denormalization can sometimes reduce joins and improve BI reporting performance.

37. What is Row-Level Security?

Row-Level Security, or RLS, restricts which records a particular user can access.
For an enterprise dashboard, this can ensure that users only see the data they are authorized to view.
The research identifies RLS as an important dashboard architecture topic.

38. Would you use a live SQL connection or a static data extract for a dashboard?

The decision depends on the reporting requirement.
A live connection can provide more current information but may increase database or server load.
A static extract can improve reporting performance but introduces a data-refresh consideration.
The research specifically identifies this trade-off as a dashboard architecture discussion.

39. Which KPIs would you track on an MIS dashboard?

The research provides several examples, including:

• Operating Cash Flow
• Current Ratio
• Days Sales Outstanding
• Accounts Receivable Turnover
• Collection Effectiveness Index
• Cash Conversion Cycle
• Reporting defect rates
• Order fulfillment cycle times

The correct KPIs ultimately depend on the business process and decision the dashboard is designed to support.

40. How do you decide whether a KPI belongs on an executive dashboard?

A KPI should have a clear relationship to the business decision being supported.
The analyst should understand:

• What is being measured?
• Why does it matter?
• What action follows a change?
• What threshold requires attention?
• How reliable is the underlying data?

The research emphasizes that an MIS Analyst needs to understand not only how to measure something, but what should be measured.

41. How do you ensure data accuracy in regulatory reports?

A strong answer should discuss multiple layers of control rather than relying on a single manual review.
The research identifies:

• Automated data-quality checks
• Anomaly detection
• Validation workflows
• Reference-data controls
• Maker-checker processes
• Accurate and complete reporting

as important elements of a regulatory reporting control framework.

42. What is SOX compliance?

The Sarbanes-Oxley Act establishes requirements around internal controls over financial reporting for public companies.
For an MIS Analyst, the relevant issue is how systems and reports support reliable financial information and auditable controls.
The research specifically connects MIS systems with Information Produced by Entity (IPE) used by external auditors.

43. What controls would you implement to ensure accurate financial reporting?

A strong answer can include:

• Segregation of Duties
• Restricted privileged access
• Immutable access logs
• Documentation of data lineage
• Validation controls
• Audit trails
• Controlled changes

The research gives the example that one user should not have unrestricted ability both to create a vendor record and approve payment to that vendor.

44. What is data lineage, and why is it important?

Data lineage describes how data moves through systems and transformations from its source to its final report or output.
It is important because analysts and auditors need to understand where reported information originated, what transformations were applied, and how the final figure was produced.
The research identifies end-to-end data-lineage documentation as part of strong financial reporting controls.

45. How would you detect an error in regulatory reporting data?

Begin with automated validation and data-quality rules.
Potential checks include:

• Missing fields
• Invalid identifiers
• Unexpected values
• Duplicates
• Broken reference-data relationships
• Unusual changes in volume
• Failed reconciliation checks

The research specifically mentions anomaly detection and malformed or missing Legal Entity Identifiers as examples.

46. Why is data governance especially important in banking?

Banking data can influence financial reporting, risk management, regulatory submissions, and operational decisions.
The research uses historical regulatory enforcement to illustrate the consequences of inaccurate or poorly controlled reporting. It describes a 2019 FCA fine against Goldman Sachs International related to inaccurate and untimely transaction reporting involving more than 220 million transactions.

47. Why do you want to join Goldman Sachs?

Avoid relying only on broad statements about prestige or brand recognition.
The research recommends connecting a specific fact about the relevant Goldman Sachs division to the candidate's own experience and then explaining the candidate's learning motivation.
For example, an MIS candidate could connect experience in ETL automation with the division's data and technology environment, provided the example genuinely reflects their background.

48. Tell me about a time you disagreed with a senior stakeholder.

Use the STAR framework:

• Situation: Explain the context.
• Task: Define your responsibility.
• Action: Explain what you did.
• Result: Explain what changed.

For an MIS or control role, the research emphasizes objective, data-driven communication and maintaining appropriate controls even when working with senior stakeholders.

49. Tell me about a significant mistake or failure you made.

A strong response should demonstrate:

• Ownership
• Immediate mitigation
• Transparent communication
• Root-cause analysis
• A preventive control or process improvement

The research specifically emphasizes that candidates should explain what systemic or automated control they introduced to reduce the likelihood of the same error happening again.

50. What would you do if a senior stakeholder asked you to bypass a data validation rule?

This question tests judgment, control awareness, and communication.
A strong response should explain that the candidate would:

• Understand the urgency
• Identify the specific control and associated risk
• Avoid bypassing an important control simply for convenience
• Explain the consequences objectively
• Escalate through the appropriate process when necessary
• Seek a compliant solution

The research frames this type of scenario around the need for MIS and control professionals to challenge inappropriate requests using evidence, established policies, and risk considerations.