LeetCode Goldman Sachs Interview — Coding + Quant Prep (2026)

Goldman Sachs is best known as the world's premier investment bank, but its Technology Division is one of Wall Street's largest engineering employers with over 12,000 engineers building trading systems, risk platforms, risk analytics, and consumer products. When candidates picture a finance interview, they imagine light SQL questions and soft behavioral rounds. Goldman Sachs is a different animal entirely.

The Goldman Sachs engineering interview is consistently rated by candidates as approaching FAANG-lite difficulty. Unlike JPMorgan or Bank of America, where a solid LeetCode Easy/Medium foundation can carry you, Goldman expects strong command of LeetCode Mediums and the occasional Hard, combined with system design fundamentals and object-oriented design fluency. This is especially true for the Engineering division, the Marcus consumer banking team, and Quantitative Engineering roles.

Goldman Sachs engineering interviews are consistently rated 1 difficulty tier above other major banks and approximately equal to non-FAANG tech companies like Stripe or Airbnb. That benchmark sets expectations clearly: if you prep as you would for a typical bank, you will underperform. If you prep as you would for Stripe or Airbnb, you will be competitive.

This guide walks through the full Goldman Sachs interview pipeline — from online assessment to final CodePair rounds — and gives you a precise problem pattern breakdown, the SQL and financial systems knowledge that shows up in data engineering tracks, a firm-by-firm comparison of finance tech interviews, and a 6–8 week prep plan to get you interview-ready.

The Goldman Sachs interview process has four main stages: application review, an online assessment (HackerRank OA), one or two CodePair phone screening rounds, and a final Superday that includes more coding, system design, and behavioral interviews. The exact format varies slightly by role (SDE, Quantitative Engineering, Technology Analyst) and seniority, but the core pattern holds across most engineering hires.

The HackerRank online assessment is 2–3 algorithmic coding questions with a 90-minute time limit. The difficulty is typically 1 Medium and 1 Medium-Hard. Partial credit is awarded for passing more test cases, so always write the brute-force solution first and optimize if time allows. Some OAs also include multiple-choice questions on CS fundamentals — time complexity, data structures, and basic OS/networking.

Goldman Sachs Technology Division uses CodePair for real-time coding interviews — communication and thinking-aloud skills are as important as the final solution. Interviewers watch you code live and expect a running dialogue. They want to understand your thought process: why you chose a particular data structure, what edge cases you are considering, how you would optimize if given more time.

New grad Technology Analyst interviews typically run 4–6 weeks from OA to offer. Experienced SDE hires move faster — 2–4 weeks once the screen is scheduled. The Superday for full-time roles includes 3–5 rounds in a single day: 2 coding rounds, 1 system design round, and 1–2 behavioral rounds with senior engineers or VPs. For internship positions, the Superday is typically 2 coding rounds plus a brief behavioral conversation.

The behavioral component at Goldman Sachs draws heavily on the firm's core competencies — client focus, teamwork, integrity, and excellence. Prepare 3–4 STAR stories that highlight collaboration on technically complex problems, situations where you balanced technical debt with delivery timelines, and examples of quantifiable technical impact. Goldman interviewers are engineers, so your behavioral stories should include technical specifics rather than generic project descriptions.

Goldman Sachs coding rounds center on LeetCode Medium difficulty with occasional Hard problems appearing in senior SDE and Quantitative Engineering interviews. The problem space leans toward classic algorithmic topics: string manipulation, tree and graph traversal, dynamic programming, and array/matrix problems. Unlike some FAANG companies that favor novel variations, Goldman tends to use cleaner problem formulations where pattern recognition and clean implementation matter most.

String problems at Goldman frequently involve parsing structured data — stock ticker symbols, financial identifiers (CUSIP, ISIN), or log formats. These problems test whether you can cleanly handle edge cases (empty strings, malformed inputs, boundary conditions) while maintaining readable code. Interviewers specifically note that Goldman engineers value code readability and structure, not just correctness.

Tree and graph problems are the second most common category. Binary tree traversals, lowest common ancestor, path sum variations, and level-order traversal all appear regularly. Graph problems tend toward BFS/DFS reachability — problems that map naturally to financial concepts like dependency resolution in trade settlement pipelines or network connectivity in system topology.

Dynamic programming questions at Goldman are typically 1D or 2D DP — house robber variants, coin change, longest common subsequence, and edit distance. Goldman is less likely than Google or Meta to ask intricate 3D DP or bitmask DP. Focus your DP prep on the 15–20 most canonical patterns rather than exotic variations.

Object-oriented design questions appear in some Goldman rounds, particularly for roles touching the firm's core systems. Common prompts include designing an order book (a core market microstructure data structure), a portfolio manager, or a rate limiter. These questions test whether you can translate a real financial concept into clean class hierarchies with appropriate encapsulation and extension points.

Time management in CodePair is critical. Goldman interviewers report that many candidates spend too long on the optimal solution when a working brute-force would have been sufficient. Always aim to have a running solution with correct logic within the first 25–30 minutes, then optimize. A clean O(n^2) solution beats a buggy O(n log n) solution every time.

Based on candidate reports and the nature of Goldman's systems, eight problem patterns appear most frequently in Goldman Sachs coding interviews. Mastering these patterns — not just the individual problems — gives you the flexibility to handle novel variations that appear in live CodePair sessions.

Divide and conquer is a Goldman favorite because it maps to how financial data is naturally processed at scale. Problems like merge sort variants, finding the k-th largest element, and binary search on answer are canonical. The key insight for Goldman contexts is that divide and conquer often models how large datasets are partitioned across trading systems and aggregated in reporting pipelines.

Two-pointer and sliding window patterns appear in string and array problems alike. Goldman interviewers particularly like problems where the optimal window or pointer movement is non-obvious — such as longest substring without repeating characters, minimum window substring, and container with most water. Practice the template until pointer initialization and termination conditions are muscle memory.

Hash map problems at Goldman often involve frequency counting, grouping anagrams, or two-sum-family lookups. The financial flavor: grouping trade records by instrument type, counting duplicate order IDs, or efficiently looking up counterparty exposure. These are Medium-difficulty but Goldman expects O(n) solutions with clear justification for your choice of hash map vs array.

Tree traversal problems (in-order, pre-order, post-order, level-order, and ZigZag) are tested in both recursive and iterative forms. Goldman interviewers often ask for the iterative version after you give the recursive solution — be ready to implement BFS with a queue or DFS with an explicit stack.

For DP, focus on the top 15 patterns: Fibonacci family, 0/1 knapsack, unbounded knapsack, longest common subsequence, edit distance, matrix DP, and interval DP. Goldman problems in this category include coin change, house robber, unique paths, and word break.

Graph problems — BFS/DFS, topological sort, union-find — round out the preparation. Number of Islands, Course Schedule, and Clone Graph are all fair game. For senior roles, expect a Dijkstra or minimum spanning tree problem framed around network routing or dependency resolution.

Goldman Sachs data engineering, quantitative engineering, and some full-stack SDE tracks include SQL rounds alongside algorithmic coding. The SQL questions are notably more sophisticated than what you encounter at typical tech companies — they reflect the financial domain Goldman operates in: time series trade data, portfolio aggregation, and risk reporting.

Common SQL topics at Goldman include window functions (RANK, DENSE_RANK, LAG, LEAD), CTEs for multi-step aggregation, self-joins for comparing records within the same table, and date arithmetic for financial period calculations. A typical Goldman SQL prompt might ask you to compute the rolling 7-day average daily PnL per trading desk, identify the top 3 performing instruments in each asset class by quarter, or find all trades where the execution price deviated from the arrival price by more than 2 basis points.

Understanding basic financial terminology gives you an edge in these rounds. Knowing that PnL means profit and loss, that a basis point is 0.01%, that trade settlement involves T+2 date arithmetic, and that an order book has bid and ask sides will help you parse Goldman SQL prompts faster and ask sharper clarifying questions.

For data engineering roles, Goldman sometimes asks about time series database design — how you would store and query tick-by-tick price data, how partitioning and indexing strategies affect latency for financial analytics, and how you would handle late-arriving data in a stream processing pipeline. These questions blend SQL knowledge with systems thinking and are a good proxy for the kind of work Goldman's Strats (quantitative strategists) and technology teams do daily.

Even if your target role is not a data engineering position, brushing up on window functions and CTEs is worthwhile. Goldman's backend systems are heavily data-driven, and demonstrating SQL fluency signals that you understand the data layer your code will interact with — a quality Goldman interviewers specifically look for in candidates.

Finance tech interviews are not monolithic. Goldman Sachs, Citadel, and JPMorgan represent three distinct points on the difficulty and style spectrum. Understanding where each firm sits helps you calibrate prep intensity and focus correctly.

Goldman Sachs sits at the top of traditional bank difficulty. As covered throughout this guide, expect LeetCode Mediums and occasional Hards, OOP design questions, system design, and strong behavioral rigor. Goldman interviewers are experienced engineers who probe your thought process carefully. The bar is closer to a scaled tech company like Stripe or Airbnb than to a typical bank.

Citadel and its market-making arm Citadel Securities are a step harder still on the quantitative side. If you are applying for a quant developer or software engineer role at Citadel, expect LeetCode Hards, advanced system design with latency and throughput constraints specific to high-frequency trading, and probability/statistics questions alongside coding. Citadel recruiters are explicit that they want candidates who understand both algorithms and market microstructure.

JPMorgan Chase sits a tier below Goldman in coding difficulty. JPMorgan SDE and technology analyst interviews typically feature LeetCode Easy-to-Medium problems, lighter system design requirements, and a heavier emphasis on behavioral questions about teamwork, leadership, and business impact. The online assessment and coding rounds are more approachable, making JPMorgan a reasonable target for candidates building toward Goldman or Citadel-level interviews.

The practical implication for prep strategy: if Goldman Sachs is your target, your preparation intensity should mirror what you would do for Stripe, Robinhood, or DoorDash — not what you would do for a typical bank. If Citadel is the target, add statistics and probability, advanced concurrency, and low-latency systems design on top of your Goldman prep. JPMorgan preparation is a solid foundation to build from but is insufficient on its own for Goldman-level rounds.

A 6–8 week focused preparation timeline is sufficient for most candidates with a solid CS fundamentals base. The key is disciplined pattern-based studying rather than random problem grinding. Goldman Sachs interviews reward depth of pattern knowledge over breadth of problem count — knowing 8 patterns cold will serve you better than having glanced at 300 problems.

Weeks 1–2: Build the foundation. Cover the core algorithmic patterns — two pointers, sliding window, hash maps, binary search, and basic tree traversal. Solve 5–7 representative LeetCode Mediums per pattern until you can recognize the pattern from the problem statement within 2–3 minutes. Use YeetCode flashcards to drill time and space complexity for each pattern.

Weeks 3–4: Advance to harder patterns. Cover dynamic programming (1D and 2D), graph BFS/DFS, topological sort, and union-find. Add 1–2 LeetCode Hards per pattern to stress-test your understanding. Begin OOP design practice — implement an order book, a rate limiter, and a simple portfolio manager in your language of choice.

Weeks 5–6: Integrate and simulate. Do timed 45-minute CodePair simulations covering 1–2 problems per session. Practice thinking aloud as you code — narrate your approach, trade-offs, and edge cases. Add system design prep: design a trade execution system, a market data feed aggregator, or a fraud detection pipeline using standard distributed systems components.

Weeks 7–8 (if available): Polish and target. Review every problem you got wrong in simulations. Do mock interviews with a peer or on a platform like Pramp. Revisit the SQL window function patterns and the financial domain terminology covered in Section 5. Read Goldman's recent engineering blog posts to have genuine talking points in behavioral rounds about the firm's technical direction.

Goldman Sachs offers strong compensation, world-class infrastructure exposure, and the prestige of working on systems that process trillions of dollars in daily trading volume. The interview is hard precisely because the role is demanding. Candidates who approach it with the same seriousness they would give a FAANG interview — disciplined pattern study, live coding practice, and system design depth — consistently succeed.

YeetCode's structured flashcard system and curated Goldman Sachs problem set can accelerate your prep significantly. The spaced repetition engine ensures you retain patterns across the full 6–8 weeks rather than cramming and forgetting. Start your prep plan today and hit the interview with confidence.