Interview experience
Goldman Sachs Interview Questions and Answers (2026)
Overview
Section titled “Overview”Goldman Sachs runs a HackerRank-plus-Superday loop where standard DSA and CS-fundamentals questions are layered with finance context and division-specific fit checks.
Goldman Sachs interview process at a glance
Section titled “Goldman Sachs interview process at a glance”| Round | Duration | What they test |
|---|---|---|
| HireVue Video Interview | ~30 min | Recorded answers to behavioral prompts |
| HackerRank Assessment (tech roles) | 90-120 min | 2-3 coding problems, aptitude/reasoning MCQs |
| Phone/Virtual Interviews | 30-60 min each | DSA, CS fundamentals, project deep-dive, light finance context |
| Superday | 2-5 back-to-back 30-60 min interviews | Technical depth, behavioral fit, division alignment |
HireVue Video Interview
Section titled “HireVue Video Interview”A recorded, one-way video interview where you answer pre-set behavioral prompts on camera with no live interviewer. It’s an early filter before any technical evaluation, so tone, structure, and clarity matter as much as content.
Common questions
- Why Goldman Sachs, and why this specific division?
- Tell me about a time you worked under pressure or a tight deadline
- Describe a leadership or teamwork experience
- Do you have any ethical concerns about a career in finance?
HackerRank Assessment
Section titled “HackerRank Assessment”For engineering applicants, a timed online test mixing 2-3 coding problems with 10-15 quantitative/logical-reasoning and data-interpretation MCQs. Coding problems tend to sit at medium difficulty.
Common questions
- Stock Buy and Sell with multiple transactions (valley-peak / greedy approach)
- Longest Increasing Subsequence
- Numerical reasoning, logical reasoning, and data-interpretation MCQs
- Array/string manipulation problems
Technical phone/virtual interviews
Section titled “Technical phone/virtual interviews”One or two rounds with an analyst, VP, or Executive Director mixing live DSA coding, a resume project deep-dive, CS fundamentals, and often a finance-flavored question or classic puzzle.
Common questions
- Find Median from Data Stream (two-heap approach) / design an LRU Cache
- Design a rate limiter (token bucket) or sketch a trading system’s order matching
- Course Schedule / cycle detection via topological sort (Kahn’s algorithm)
- Explain database ACID properties, eventual consistency, and TCP vs UDP
- Classic logic puzzle - e.g. find the heavier of 8 balls in the minimum weighings
- “Explain how stock markets work” or “what is an option?” at a basic level
Full round-by-round narratives are on the Goldman Sachs interview experience page.
Superday
Section titled “Superday”Goldman Sachs’s final round: 2-5 back-to-back 30-60 minute interviews in a single sitting, in person or over video, with analysts, associates, and sometimes VPs or Executive Directors. Each interviewer runs an independent technical-plus-behavioral evaluation, so expect the same ground - live coding, project discussion, “why Goldman Sachs” - to be covered more than once in one day.
Common questions
- Walk me through your most complex resume project, end to end
- Live coding problem with the interviewer watching your approach, not just the answer
- Why this division specifically, and how does it fit your background?
- Tell me about a leadership experience or a time you handled a disagreement in a team
Sample answer frameworks are on the Goldman Sachs HR interview questions page.
Common technical interview questions and answers
Section titled “Common technical interview questions and answers”Q: How do you solve Best Time to Buy and Sell Stock with multiple transactions?
When unlimited transactions are allowed, the answer is the sum of every positive consecutive difference: for each day i, if price[i] is greater than price[i-1], add the difference. This valley-peak greedy is O(n) time and O(1) space, and it is provably optimal because any multi-day gain decomposes into consecutive daily gains. The single-transaction variant instead tracks the minimum price seen so far and the best profit measured against it, also O(n). The at-most-k-transactions variant needs DP over buy[j] and sell[j] states for j from 1 to k, giving O(nk) time.
Q: How do you find the median from a data stream?
Maintain two heaps: a max-heap holding the lower half of the values and a min-heap holding the upper half. On each insert, push into one heap and then rebalance so their sizes differ by at most one, moving the offending root across. The median is the root of the larger heap, or the average of the two roots when the sizes are equal. Insertion costs O(log n) and reading the median is O(1). A common follow-up asks what changes if the values are bounded small integers - then a counting array with a running prefix, or an order-statistic tree, becomes the better structure.
Q: How do you implement an LRU cache?
Combine a hash map with a doubly linked list. The map gives O(1) lookup from key to node, and the list keeps nodes in recency order with the most recently used at the head. On get, move the node to the head; on put, insert at the head and, if capacity is exceeded, remove the tail node and delete its key from the map. Using sentinel head and tail nodes removes the null-pointer edge cases. Both operations are O(1). In Java you can shortcut with LinkedHashMap in access order and an overridden removeEldestEntry, but interviewers usually want the manual construction.
Q: How does topological sort solve the Course Schedule problem?
Model courses as nodes and prerequisites as directed edges, then run Kahn’s algorithm: compute in-degrees, enqueue every node with in-degree zero, and repeatedly pop a node, append it to the order, and decrement its neighbours’ in-degrees, enqueuing any that reach zero. If the resulting order holds fewer nodes than the graph has, a cycle exists and the schedule is impossible. Complexity is O(V + E) time and O(V) space. The DFS alternative colours nodes white, grey, and black, reports a cycle on reaching a grey node, and produces the order by reverse post-order.
Q: How would you design a rate limiter?
The token bucket is the standard answer: a bucket holds up to N tokens, refills at a fixed rate, and each request consumes one token, being rejected with HTTP 429 when the bucket is empty. It permits short bursts up to the bucket size while capping the long-run rate, which is why public APIs favour it. Fixed-window counters are simpler but allow twice the limit through at a window boundary; a sliding-window log is exact but stores every timestamp; the sliding-window counter is the practical compromise. Distributed enforcement needs shared state in Redis with an atomic INCR-and-expire or a Lua script, so two nodes cannot both admit the same request.
Q: What are the ACID properties of a database transaction?
Atomicity means a transaction either fully commits or fully rolls back, implemented through the write-ahead log and undo records. Consistency means the transaction moves the database from one valid state to another, respecting constraints, keys, and triggers. Isolation means concurrent transactions do not observe each other’s intermediate state, tuned by isolation level - read committed avoids dirty reads, repeatable read avoids non-repeatable reads, and serializable also avoids phantoms, achieved through locking or MVCC snapshots. Durability means a committed transaction survives a crash, guaranteed by flushing the log to stable storage before the commit is acknowledged.
Q: What is the difference between TCP and UDP, and where does each fit in a trading system?
TCP is connection-oriented: it performs a three-way handshake, numbers bytes, retransmits losses, reorders segments, and applies flow and congestion control, so delivery is reliable and ordered but latency varies and head-of-line blocking is possible. UDP is a thin datagram wrapper over IP with no handshake, retransmission, or ordering, so it carries far less overhead and far more predictable latency. Order entry and execution confirmations use TCP, because a dropped order is unacceptable. High-volume market-data feeds typically use UDP multicast, since one packet reaches every subscriber and a slightly stale quote beats a delayed one, with gaps recovered over a separate retransmission channel.
Q: Explain what an option is and how stock markets work at a basic level.
An exchange matches buyers and sellers through a central limit order book, prioritising resting limit orders by price and then by arrival time; a market order crosses the spread and executes against the best available opposite orders. Prices move as order flow consumes liquidity on one side of the book. An option is a contract giving the holder the right, but not the obligation, to buy (a call) or sell (a put) the underlying at a fixed strike price on or before expiry, in exchange for a premium paid to the writer. A call has intrinsic value when the spot price is above the strike, and the premium is intrinsic value plus time value, which grows with remaining time and volatility. Engineering candidates are not expected to price options - only to hold this conversation confidently.
Frequently asked questions about Goldman Sachs interviews
Section titled “Frequently asked questions about Goldman Sachs interviews”What is the Goldman Sachs interview process for freshers?
Goldman Sachs typically runs 3-4 stages: 1. A HireVue recorded video interview (about 30 minutes) where you answer pre-set behavioral prompts on camera. 2. A HackerRank coding assessment for engineering applicants (60-120 minutes) covering 2-3 coding problems plus aptitude/reasoning MCQs. 3. Phone/virtual interviews (around 30-60 minutes each) with analysts and VPs from the specific division you applied to, mixing DSA, project deep-dives, and light finance questions. 4. The Superday - the final round, with 2-5 back-to-back 30-60 minute live interviews with analysts, associates, and sometimes VPs or Executive Directors in one sitting, testing technical depth, behavioral fit, and division alignment together.
What questions are asked in Goldman Sachs interviews?
Technical rounds mix classic DSA (stock buy/sell variants, longest increasing subsequence, LRU cache, median from a data stream, topological sort for course scheduling), CS fundamentals (ACID properties, eventual consistency, TCP vs UDP), and lightweight system-design prompts like designing a rate limiter or a trading system. Expect at least one finance-flavored question - explain how stock markets work, what an option is - and a classic logic puzzle (the 8-balls weighing problem shows up often). Behavioral rounds probe teamwork, pressure situations, and why you want that specific division at Goldman Sachs rather than banking in general.
How many rounds are there in the Goldman Sachs interview?
Most candidates go through a HireVue video interview, a HackerRank OA (for tech roles), one or two technical phone/virtual interviews, and finally a Superday - a single day of 2 to 5 back-to-back interviews with people across the division. The Superday is where most final offer decisions get made; candidate reports describe 3-4 stages total from OA to Superday over roughly 2-4 weeks.
What is the Goldman Sachs Superday like?
Superday is Goldman Sachs’s final-round format: several 30-60 minute interviews back-to-back in one day (in person or over video), typically with an analyst, an associate, and a more senior VP or Executive Director. Each interview restarts the technical and behavioral evaluation from scratch, so candidates report being asked to code live, discuss a resume project in depth, solve a puzzle, and answer ‘why Goldman Sachs’ more than once in the same day. Staying consistent and high-energy across every interview matters as much as any single answer.
How should I prepare for Goldman Sachs interviews?
Practice 150-200+ DSA problems at medium difficulty (arrays, heaps, graphs, DP), revise CS fundamentals like database ACID properties and networking basics, and learn enough finance to talk comfortably about stocks, options, and risk at a beginner level. Rehearse the HireVue format by recording yourself answering common behavioral prompts, prepare a sharp, division-specific answer for ‘why Goldman Sachs’ and ‘why this division’, and be ready to repeat clear, structured answers across several back-to-back Superday interviews without losing energy.
Is Goldman Sachs’s technical interview different from a typical product company’s?
Mostly no for the coding bar - it’s standard DSA plus CS fundamentals, similar to other large tech recruiters. The difference is the finance layer stacked on top: interviewers weave in market/trading context (explaining a trading system’s order matching and latency needs, or a basic options question) even in engineering loops, and the closing HR round often checks for genuine interest in the finance industry, not just software engineering in the abstract.