Goldman Sachs uses HackerRank to assess applicants for Engineering roles. According to Goldman's own guide, published in August 2026, every candidate takes a 30-minute Software Fundamentals challenge of data structures and algorithms problems. Engineering applicants add a 45-minute Software Diagnostics challenge, fixing issues in an existing Django, Node.js or Spring Boot app; Quantitative Strategist applicants add a 22-minute multiple-choice Math challenge. Prepare with timed algorithm practice plus hands-on debugging in unfamiliar code, and complete the assessment on your own under Goldman's rules.
Much of what still circulates describes an older, longer multi-section test. This guide sticks to what Goldman currently publishes and marks where candidate reports fill the gaps.
The format at a glance
Goldman says candidates are assigned challenge types based on their role preferences. Here is how the pieces fit together:
| Challenge | Who takes it | Time | What it covers |
|---|---|---|---|
| Software Fundamentals | All candidates | 30 min | Data structures and algorithms problems in the language of your choice; time and space complexity |
| Software Diagnostics | Engineering applicants | 45 min | Diagnosing and resolving issues in an existing app built with Django (Python), Node.js (JavaScript) or Spring Boot (Java) |
| Math | Quantitative Strategist applicants | 22 min | Multiple-choice questions on probability, statistics, calculus and linear algebra |
Add it up and an Engineering applicant faces about 75 minutes of challenge time (30 + 45), while a Quantitative Strategist applicant faces about 52 (30 + 22). Your invitation is the final word on which challenges you get.
Two caveats. First, this guidance comes from Goldman's Engineering Campus Recruiting team, so it describes campus (internship and new-grad) hiring. Experienced-hire processes may differ; ask your recruiter whether there is an assessment and what it contains. Second, do not study from older write-ups. A 2021 candidate report on GeeksforGeeks, for example, described a 135-minute HackerRank test with five mandatory sections, including computer science and quantitative multiple-choice questions and a short written section. Goldman's current guide describes nothing like that.
Software Fundamentals: the 30-minute coding challenge
Goldman describes this challenge as data structures and algorithms problems solved in the programming language of your choice, and tells candidates to review core concepts: data structures, algorithms, and time and space complexity. It does not publish the number of problems or their difficulty, so expect both to vary.
How HackerRank scores your code
HackerRank's documentation explains that each coding question runs against sample test cases, which you can see, and hidden test cases, which you cannot. The score for a question is the sum of the points for every test case you pass, and HackerRank advises question authors to give harder test cases more points. In practice that means:
- Passing the samples is not the finish line. The points that matter most usually sit in the hidden cases, so test empty input, a single element, duplicates, negative numbers and the largest sizes the constraints allow.
- Partial credit is real. A correct but slow solution can still earn the cases it finishes in time. Get something working before chasing the optimal version.
- Constraints tell you the target complexity. If n can reach 105 or more, a nested loop over the input is unlikely to pass every case; reach for sorting, hashing or a single pass.
For more on the platform itself, from the editor to custom input, see our HackerRank assessment guide.
What the problems tend to look like
The 2021 report mentioned above described two programming questions in a 30-minute coding section, but Goldman has not said whether today's challenge is similar. Plan for standard interview problems on the topics Goldman lists, in roughly this order:
- Arrays and hashing — counting, grouping, complement lookups, prefix sums.
- Strings — parsing, frequency maps, sliding windows.
- Sorting and two pointers — pair finding, merging and interval problems.
- Stacks and queues — bracket matching, monotonic stacks, simulations.
- Trees and graphs — recursive traversal, BFS on grids, connected components.
- Dynamic programming — one-dimensional recurrences and simple grid DP.
For worked examples in this style, see our Goldman Sachs coding interview questions guide.
A 30-minute game plan
- Read everything first (2–3 minutes). Note the constraints and the input format for each problem.
- Start with the problem you are surest of. Banking points early beats getting stuck on the hardest one.
- Write the simplest correct version, then run the samples.
- Test your own edge cases with custom input before you move on.
- Optimize only if the constraints demand it, and keep a working version you can fall back to.
- Watch the clock. Leave a few minutes at the end to make sure your best working code is what gets submitted.
Software Diagnostics: the 45-minute debugging challenge
This is the section that pure LeetCode practice does not cover. Goldman says Engineering candidates work inside an existing application built with Django (Python), Node.js (JavaScript) or Spring Boot (Java) to diagnose and resolve issues in a realistic engineering environment. Its prep advice is specific: practice working in an existing codebase, with the focus on understanding unfamiliar code, debugging, writing tests and making targeted changes rather than building something from scratch.
Goldman's guide also notes that this challenge includes a built-in, "limited" AI assistant inside the test environment. It will not give you the full solution; it can help with navigating the interface, syntax or clarifying the problem statement, and Goldman expects you to apply your own engineering judgment to evaluate and validate the final fix. That tool is part of Goldman's test design. Anything outside the test environment is not: follow the on-screen rules and do the work on your own. Either way, the diagnosis and the fix have to come from you, so drill the debugging loop below until it is automatic.
The debugging loop to practice
- Orient. Skim the project layout: where routes or controllers live, where the data models are, and where the tests are.
- Reproduce. Run the failing test or request and read the whole error and stack trace, not just the last line.
- Trace. Follow the path from the route or controller to the business logic to the data layer.
- Narrow it down. Form one hypothesis at a time and confirm it with a log line or a focused test.
- Fix the root cause minimally. Change the smallest thing that corrects the behavior and resist refactoring.
- Verify. Run the full test suite, not only the test you were chasing, and add a regression test where it makes sense.
Know your way around all three stacks
Goldman lists three frameworks but does not say how the stack is assigned, so get fluent in the one closest to your main language and learn to read the other two. These are standard framework conventions, not details of Goldman's app:
| Stack | Where the logic usually lives | How tests usually run |
|---|---|---|
| Django (Python) | urls.py → views.py → models.py, plus serializers if it uses Django REST Framework | python manage.py test, or pytest if configured |
| Node.js (JavaScript) | Route files → controllers or handlers → services and data access | Whatever the test script in package.json runs, usually via npm test |
| Spring Boot (Java) | @RestController → @Service → @Repository and entities | mvn test or ./gradlew test |
A good drill: clone a small open-source app in your main stack, ask a friend to break something (or revert a bug-fix commit from its history), and give yourself 45 minutes to find it, fix it and get the suite green. Bugs worth practicing on include off-by-one errors in pagination, missing null or empty checks, wrong HTTP status codes, incorrect query filters, date and time-zone handling, and in Node, a promise that is never awaited.
The Math challenge (Quantitative Strategist roles)
Strats applicants get a 22-minute multiple-choice section covering probability, statistics, calculus and linear algebra. That is a short clock, so speed on fundamentals matters more than depth:
- Probability: conditional probability, Bayes' rule, expected value and variance, common distributions.
- Statistics: mean, variance, standard deviation and basic inference.
- Calculus: derivatives, integrals and single-variable optimization.
- Linear algebra: matrix multiplication, determinants, inverses and eigenvalues.
Goldman does not publish how this section is scored, so read the on-screen instructions before deciding whether to guess on questions you cannot finish.
Rules, proctoring and test-day setup
Complete the assessment on your own, under Goldman's rules and the instructions in your invitation. Goldman does not publish the monitoring settings for its tests, but HackerRank's documentation shows what employers can switch on:
- On by default: tab proctoring, which records how often and how long you leave the test window, and copy-paste tracking on coding questions.
- Secure Mode: enforces full screen, blocks copy and paste, prevents multiple monitors and flags tab switches.
- Proctor Mode: adds screenshot analysis, plagiarism detection and webcam anomaly detection.
- Desktop App Mode: adds operating-system-level monitoring through a native app.
Read the test instructions and any proctoring consent screen before you start; they tell you what applies to your test. Then set yourself up properly:
- Take the HackerRank practice assessment linked from Goldman's guide so the editor and test runner are not new on the day.
- Pick your language in advance and know its standard library cold: sorting with a custom key, hash maps, heaps and deques.
- Use a stable connection and a charged laptop, close other apps and turn off notifications.
- Block about two hours of quiet time even though the challenges add up to less, so a slow start does not eat into your clock.
For how other employers run their online assessments, see our new grad online assessment guide.
What happens after the assessment
Goldman says assessment results are evaluated alongside the rest of your application, and campus candidates should receive an update from the recruiting team within three weeks. You can check your status in the candidate portal at any time. Goldman's student recruiting page lists the HackerRank and a video interview together as its first-interview step, without saying which comes first, followed by a Superday:
| Stage | What Goldman says | What candidates commonly report |
|---|---|---|
| Video interview | About 30 minutes; the page refers to it as a HireVue interview and says Engineering applicants are also asked to take the HackerRank assessment | Recorded behavioral and motivation questions, such as why Goldman and why this division |
| Superday | For candidates selected after the video interview: a series of final-round interviews, typically two to five for campus hires depending on the division, with a cross-section of people you could work with | Live coding on data structures and algorithms, a deep dive on your résumé and projects, behavioral and fit questions, and in some reports probability puzzles or core computer science questions such as how a hash map works |
Both stages are spoken, so rehearse out loud rather than in your head; our HireVue software engineer questions guide covers the recorded format. One way to practice is a free AI mock interview in CoPilot Interview: paste the Goldman job description and your résumé, choose a mix of behavioral and coding questions, answer aloud, and get a score on each answer with what worked and what to improve.
A three-week prep plan
| Week | Software Fundamentals | Software Diagnostics | Everything else |
|---|---|---|---|
| 1 | Arrays, hashing, strings, two pointers, sorting. Solve in 30-minute timed blocks and state the Big-O of every solution. | Set up your main stack locally; read one small open-source app end to end and run its tests. | Take Goldman's HackerRank practice assessment. |
| 2 | Stacks, trees, graphs and BFS, basic DP. Write custom edge-case inputs before every submission. | Two or three 45-minute bug-fix drills: reproduce, trace, fix, run the full suite, add a test. | Strats applicants: a short daily set of timed probability and linear algebra questions. |
| 3 | Mixed timed sets that mimic the 30-minute window; re-solve every problem you missed. | One read-through drill in each of the other two stacks. | Draft your why-Goldman and why-this-division answers plus two project stories; rehearse them aloud. |
If you have less time, keep the structure and shrink it: one timed coding block and one debugging drill per day is a better mix than coding alone.
Mistakes that cost candidates
- Preparing only with LeetCode. For Engineering applicants, Diagnostics is 45 of roughly 75 minutes of challenge time. Practice it.
- Stopping at the sample tests. Hidden test cases usually carry most of the points, so check edge cases and large inputs yourself.
- Rewriting instead of fixing. In the Diagnostics app, a small targeted change that keeps the existing tests green beats an ambitious refactor.
- Forgetting what comes next. The video interview and the Superday are where the rest of the process happens, so start your behavioral stories before the invite arrives.
Get ready for the rounds after the HackerRank
The video interview and Superday are spoken. Paste the Goldman job description and your résumé into a free AI mock interview, answer out loud, and get a score plus what to improve on every answer, with an interview report at the end.
Try a free AI mock interviewFAQ
Does Goldman Sachs use HackerRank for engineering roles?
Yes. Goldman's careers site says HackerRank is one of the tools it uses to assess candidates for Engineering roles, and its student recruiting page says Engineering applicants will be asked to take a HackerRank assessment.
How long is the Goldman Sachs HackerRank assessment?
Goldman's 2026 guide lists 30 minutes for the Software Fundamentals challenge, 45 minutes for Software Diagnostics and 22 minutes for Math. Engineering applicants take Fundamentals plus Diagnostics, and Quantitative Strategist applicants take Fundamentals plus Math.
What is the Software Diagnostics challenge?
It is a 45-minute exercise in which you diagnose and fix issues in an existing application built with Django, Node.js or Spring Boot. Goldman recommends practicing reading unfamiliar code, debugging, writing tests and making targeted changes.
Is the Goldman Sachs HackerRank assessment proctored?
Goldman does not publish its settings, but HackerRank tracks tab switches and copy-paste by default and lets employers add full-screen, webcam and screenshot monitoring. Read the test instructions before you start and complete the test on your own under Goldman's rules.
How long does it take to hear back after the assessment?
Goldman's campus recruiting guidance says candidates should receive an update within three weeks, and you can check your status in the candidate portal.
What comes after the HackerRank assessment?
For campus candidates, Goldman lists the assessment alongside a video interview of about 30 minutes. Candidates selected after that are invited to a Superday of typically two to five interviews, depending on the division.