Quant Researcher
47 questionsInference, time series, machine learning, portfolio construction and research validation.
CH01 · 02 · 08 · 09 · 12 · 14
QuantJourney OpenBook
Quant
Quant
Interviews
◆
From Mathematical Foundations to Production-Grade Quant Systems
194 interview questions with full derivations, executable Python, C++ and SQL implementations and tests. If a numerical claim can be tested, it is.
1,630 pages · 194 questions · 2,319 rendered equations · 14 chapters in 2 volumes
An answer you can inspect and discuss.
Each question carries the derivation, implementation contract and tests behind it—so you can inspect, challenge and discuss the result, not just read the solution.
Interviewer prompt
The question as asked, with role tags and expected time.
30-second answer
The answer given first, before any derivation.
Full solution
The complete derivation; every displayed equation is numbered.
Intuition
The mechanism behind the result, in a few sentences.
Common wrong answers
Recurring mistakes, stated precisely.
Interviewer follow-ups
Where the conversation goes after a correct answer.
Implementation & production note
Estimation issues that appear when the result meets capital.
Lab — implementation contract
Inputs, invariants and failure conditions, fixed before code.
Tests & evidence
Source code and tests; listings are extracted from tested implementations.
CH01.05
Sampling Uncertainty of a Sharpe Ratio
Interviewer prompt
A daily strategy reports an annualized Sharpe ratio of 1.5 from three years of data. Estimate its uncertainty under IID Gaussian returns. Then explain how skewness, kurtosis and autocorrelation change the answer.
30-second answer
With 756 daily observations, the IID Gaussian asymptotic standard error of the annualized Sharpe is about 0.58, giving a rough 95% interval of 0.37 to 2.63. The reported 1.5 is much less precise than it looks.
Full solution — Lo (2002); Mertens (2002)
s = 1.5 / √252 ≈ 0.09449
(1.42)
SE(ŜA) ≈ √252 · ((1 + ½s²) / n)1/2 ≈ 0.579
(1.45)
CI95% ≈ [0.37, 2.63]
(1.47)
neff ≈ n (1 − ρ₁) / (1 + ρ₁)
(1.49)
Lab 1.5 — implementation contract
| Input | annualized Sharpe, observations, periods per year, skewness, excess kurtosis |
| Invariant | SE scales as inverse square root of n; rises under adverse higher moments |
| Failure | n ≤ 1, or the moment combination implies negative variance |
Tests & evidence
test_annualized_sharpe_standard_error_matches_reference_case ✓ 0.578
test_annualized_sharpe_se_scales_and_reflects_higher_moments ✓ higher moments covered
14 chapters. Search every question.
Search across all 194 question titles, concepts and chapter names. Or open a chapter to browse it in full.
No question titles match that search. Try a topic or chapter name.
Volume I — Mathematics and Markets
01Probability and Statistical Inference98 pp · 16 q
- 01Base Rates and an Apparently Strong Classifier
- 02Tower Property in a Latent-Regime Model
- 03Dependence Beyond Correlation
- 04Simpson’s Paradox in Strategy Aggregation
- 05Sampling Uncertainty of a Sharpe Ratio
- 06Block Bootstrap for Dependent Returns
- 07Maximum Likelihood and Information
- 08Robust Estimation Under Heavy Tails
- 09Five Thousand Backtests and False Discoveries
- 10Sequential Peeking and Optional Stopping
- 11Power and Minimum Detectable Alpha
- 12Winner’s Curse After Model Selection
- 13Bayesian Update for a Trading Edge
- 14Model Comparison by Predictive Evidence
- 15Heteroskedastic and Autocorrelated Regression Errors
- 16From Statistical Significance to Expected Utility
02Linear Algebra and Optimization97 pp · 14 q
- 01Why Every Covariance Matrix Is Positive Semidefinite
- 02Eigenvalues of an Equicorrelation Matrix
- 03Diagnosing an Ill-Conditioned Covariance Matrix
- 04PCA as a Risk Decomposition
- 05Factor Models and the Woodbury Identity
- 06Correlated Simulation with Cholesky and Its Failure Modes
- 07Shrinkage Covariance and Portfolio Stability
- 08Projecting an Invalid Matrix to the Nearest Correlation Matrix
- 09Markowitz Optimization with Linear Constraints
- 10Turnover-Regularized Portfolio Optimization
- 11Risk-Parity Equations and Non-Uniqueness
- 12Fast Portfolio Updates with Sherman-Morrison-Woodbury
- 13Why Solving Ax=b Beats Forming A⁻¹
- 14L1 versus L2 Constraints in Sparse Portfolio Design
03Calculus, PDEs and Numerical Methods111 pp · 12 q
- 01Newton-Raphson Convergence and Failure
- 02A Robust Implied-Volatility Root Finder
- 03Quadrature Error for Smooth and Singular Payoffs
- 04Monte Carlo versus Quasi-Monte Carlo
- 05Stability of an Explicit Finite-Difference Scheme
- 06Why Crank-Nicolson Can Oscillate
- 07Automatic Differentiation for Greeks
- 08Calibration as Weighted Nonlinear Least Squares
- 09Adjoint Sensitivities for Many Parameters
- 10Arbitrage-Aware Interpolation of Curves and Surfaces
- 11Solving a Mean-Reverting Linear ODE
- 12Boundary Conditions for a Barrier-Option PDE
04Stochastic Processes and Stochastic Calculus114 pp · 16 q
- 01Filtrations, Adaptedness and What the Trader Knows
- 02Quadratic Variation of Brownian Motion
- 03Distribution of a Deterministic Ito Integral
- 04When Is a Stochastic Integral a Martingale?
- 05Ito Lemma for the Log of an Asset Price
- 06Solving Geometric Brownian Motion
- 07Conditional Distribution of an Ornstein-Uhlenbeck Process
- 08The Ratio of Two Correlated Geometric Brownian Motions
- 09Exponential Martingales and Novikov Conditions
- 10Girsanov Change of Measure in One Factor
- 11Martingale Representation and Market Completeness
- 12Hitting Probabilities and Stopping Times
- 13Feynman-Kac from Pricing PDE to Expectation
- 14Strong and Weak Error of Euler-Maruyama
- 15When Milstein Improves on Euler
- 16Simulating Correlated Diffusions Safely
05No-Arbitrage and Derivatives118 pp · 16 q
- 01One-Period No-Arbitrage Bounds
- 02State Prices and Risk-Neutral Probabilities
- 03Put-Call Parity with Dividends and Stock Borrow
- 04Bounds on Vertical Spreads
- 05Butterfly Convexity across Strikes
- 06Calendar Spreads, Forwards and Carry
- 07Auditing a Sparse Option Surface for Static Arbitrage
- 08Replicating a Digital with Tight Call Spreads
- 09American Exercise in a Binomial Tree
- 10Risk-Neutral Pricing without Assuming Real-World Drift
- 11Deriving the Black-Scholes PDE
- 12Pricing a Forward-Start Option
- 13Quanto Drift Adjustment
- 14Barrier In-Out Parity and Monitoring Risk
- 15Dupire Local Volatility: What the Surface Implies
- 16Model-Independent Bounds for an Exotic Payoff
06Volatility, Greeks and Options Trading114 pp · 14 q
- 01Greeks by Units, Signs and Limiting Cases
- 02Delta-Hedged Option P&L and Realized Variance
- 03Constructing a Long-Gamma Short-Vega Trade
- 04Vanna and Volga Risk in a Smile Book
- 05Sticky Strike versus Sticky Delta
- 06Equity Skew and Crash-Risk Compensation
- 07Inverting Black-Scholes Implied Volatility
- 08Building a No-Arbitrage Volatility Surface
- 09Variance-Swap Replication from Options
- 10Corridor Variance and Tail Exclusion
- 11Heston Dynamics and the Shape of the Smile
- 12SABR Parameters and Smile Geometry
- 13Vol-of-Vol, Roughness and Forecast Horizons
- 14Hedging with Jumps and Discrete Rebalancing
07Fixed Income, Rates and Credit93 pp · 12 q
- 01Discount Factors, Zero Rates and Forward Rates
- 02Bootstrapping a Discount Curve
- 03Duration, Convexity and Key-Rate Risk
- 04Par Swap Rate, Annuity and DV01 Hedge
- 05Multi-Curve Discounting and Basis Risk
- 06Vasicek and Hull-White Short-Rate Models
- 07The HJM Drift Restriction
- 08Hazard Rates, Survival and Risky Bond Pricing
- 09CDS Par Spread and Hazard Bootstrap
- 10Structural versus Reduced-Form Credit Models
- 11Callable Bonds, OAS and Negative Convexity
- 12Designing a Fixed-Income Risk and P&L Explain Engine
Volume II — Research and Systems
08Time Series, Econometrics and Machine Learning121 pp · 16 q
- 01What Stationarity Means for a Trading Model
- 02Reading ACF and PACF without Pattern Matching
- 03AR(1) Half-Life and Mean-Reversion Speed
- 04Unit-Root Tests and Their Low Power
- 05Cointegration versus Correlation
- 06An Error-Correction Model for a Trading Spread
- 07Kalman Filtering a Time-Varying Hedge Ratio
- 08Regime-Switching Returns
- 09GARCH, Volatility Clustering and Forecasting
- 10Forecasting Realized Volatility
- 11Cross-Validation for Time-Series Data
- 12Purging and Embargo against Label Leakage
- 13Feature Selection under Dependence
- 14Calibrating Probabilistic Predictions
- 15Monitoring Concept Drift in Live Models
- 16Prediction Accuracy versus Economic Alpha
09Portfolio Construction, Risk and Performance134 pp · 16 q
- 01The Mean-Variance Portfolio from First Principles
- 02How Covariance Error Dominates Portfolio Weights
- 03Black-Litterman as Bayesian Portfolio Construction
- 04Risk Parity and Equal Risk Contributions
- 05Neutralizing Factors without Killing Alpha
- 06Turnover, Transaction Costs and Rebalancing
- 07Expected Shortfall as a Tail-Risk Measure
- 08Why VaR Can Fail
- 09Designing a Scenario and Stress Framework
- 10Drawdown as a Path-Dependent Risk
- 11Kelly Sizing under Parameter Uncertainty
- 12Capacity under Square-Root Market Impact
- 13Performance Attribution for a Systematic Book
- 14Sharpe, Information Ratio and Breadth
- 15Optimal Rebalancing Frequency
- 16Robust Portfolio Optimization
10Market Microstructure and Execution115 pp · 12 q
- 01Reading a Limit Order Book: Spread, Midpoint, Imbalance and Microprice
- 02Price-Time Priority, Queue Position and Fill Probability
- 03Market Order or Limit Order? Expected Cost and Adverse Selection
- 04Price Impact: Kyle Lambda, Square-Root Scaling and Capacity
- 05Almgren-Chriss Optimal Execution
- 06Execution Benchmarks: Implementation Shortfall, VWAP, TWAP and POV
- 07Smart Order Routing across Venues
- 08Hidden Liquidity, Icebergs and Dark Execution
- 09Transaction Cost Analysis without False Precision
- 10Self-Exciting Order Flow and Hawkes Processes
- 11Market Making, Inventory Risk and Avellaneda-Stoikov Quotes
- 12Designing a Production Execution Platform
11Programming, Data Structures, Python, C++ and SQL163 pp · 16 q
- 01Complexity Is Not a Latency Model: Big-O, Locality and the Amdahl Speedup Law
- 02Streaming Top-k with Deterministic Tie Handling
- 03Exact versus Approximate Rolling Quantiles
- 04Hash Maps, Symbol Tables and an LRU Cache That Does Not Lie
- 05Dependency DAGs, Topological Builds and Cycle-Safe Invalidation
- 06C++ Concurrency: False Sharing, Atomics and Bounded Queues
- 07RAII, Ownership and Exception-Safe Quant Components
- 08Floating-Point Representation, Stable Formulas and Tolerance Contracts
- 09NumPy Shapes, Broadcasting and Memory Layout
- 10Grouped Calculations: Pandas, Vectorization and SQL Parity
- 11Bitemporal Point-in-Time SQL without Look-Ahead
- 12SQL Window Functions: Lag, Ranking and Frame Semantics
- 13Price-Time Order-Book Data Structures
- 14Testing Numerical and Stochastic Quant Software
- 15Stable APIs, Schemas and Replay-Compatible Contracts
- 16Profiling, the Amdahl Speedup Law and Multi-Language Quant-System Design
12Data, Backtesting and Production System Design159 pp · 14 q
- 01Designing a Canonical Instrument Master under Identity Change
- 02Point-in-Time Data: Valid Time, Knowledge Time and As-Of Joins
- 03Corporate Actions, Adjusted Prices and Economic Return Reconstruction
- 04An Event-Driven Backtester with Explicit Clock Semantics
- 05Order Lifecycle, Partial Fills and Execution Realism
- 06Portfolio Accounting: Lots, Cash, FX, Financing and NAV
- 07Missing Data, Stale Prices, Delistings and Data-Quality Policy
- 08Reproducible Research Snapshots and Content-Addressed Lineage
- 09Walk-Forward Validation with Point-in-Time Universes and Purging
- 10Event Sourcing, Idempotency and Deterministic Replay
- 11Distributed Pipelines, Partitioned Backfills and Failure Recovery
- 12OMS, Pre-Trade Risk, Approval Rules and What-If State
- 13Reconciliation, Observability, SLOs and Incident Evidence
- 14An AI-Native Quant Platform with Deterministic Tool Boundaries
13Brainteasers, Estimation and Game Theory84 pp · 10 q
- 01Estimate Daily Equity Trading Notional from First Principles
- 02Extracting Fair Bits from an Unknown Biased Coin
- 03The Secretary Problem and Interview Timing
- 04Optimal Search with Recall, Cost and a Finite Horizon
- 05A First-Price Sealed-Bid Auction with Strategic Opponents
- 06Biased Gambler's Ruin with Unit Steps
- 07A Conditional-Probability Trap with Selection
- 08An Invariant in a Matrix Transformation Game
- 09Expected Hash Collisions in a Large System
- 10A One-Bit Communication Coordination Game
14Mock Interviews and Role Tracks109 pp · 10 q
- 01Quant Researcher Mock Interview: Signal Validation
- 02Quant Trader Mock Interview: Probability and Risk
- 03Desk Quant Mock Interview: Pricing and Hedging
- 04Quant Developer Mock Interview: Algorithms and C++
- 05Systematic PM Mock Interview: Capital Allocation
- 06Risk Quant Mock Interview: Stress and Controls
- 07ML Quant Mock Interview: Temporal Generalization
- 08Senior Platform Design Interview: Research to Execution
- 09Take-Home Research Case: From Dataset to Decision
- 10Scoring Rubric and Interviewer Calibration
Read the actual pages.
Who is this for?
Choose a role, not a reading order. These tracks cut across both volumes; the same question can appear in more than one track when the work overlaps.
Quant Trader
39 questionsDerivatives, volatility, rates, microstructure, execution and decisions under uncertainty.
CH01 · 05 · 06 · 07 · 10 · 13Quant Developer
44 questionsNumerical methods, C++, Python, SQL, data contracts, replay and production systems.
CH02 · 03 · 10 · 11 · 12 · 14Systematic PM
36 questionsEvidence quality, portfolio construction, tail risk, attribution, capacity and governance.
CH01 · 07 · 09 · 10 · 12 · 14Where most interview books stop.
These questions continue past the whiteboard into execution, C++, point-in-time data and production controls.
Market microstructure & execution
Queue position, adverse selection, market impact, smart order routing, TCA and production execution platforms.
Programming under real constraints
C++ concurrency, memory locality, numerical contracts, bitemporal SQL, stable APIs and replay-compatible components.
Research-to-production systems
PIT data, event sourcing, deterministic replay, OMS, reconciliation, observability and AI-native tool boundaries.
Not pseudocode. Executable code.
Get the code
Clone the complete repository. It contains 210 Python files (48,408 lines), 44 C++ files (4,334 lines) and 53 SQL files (5,717 lines), plus 67 public code-only test and validator files. The implementations, tests and rendered listings come from the same source tree.
Browse the repository- 194
- questions
- 58k+
- lines of code
- Python / C++ / SQL
- three languages
- Tests
- included
Text — CC BY-NC-ND 4.0 · code & tool contracts — MIT
Get the book in PDF or buy the printed edition.
Enter your email once to receive the complete PDF and all 14 individual chapter files. This is a one-time delivery, not a marketing subscription. QuantJourney research & updates remain a separate, optional choice.
Full PDF. Both volumes and all fourteen chapters.
Get the PDFTwo-volume physical edition, printed for long-form study at the desk.
Pre-order with StripePrint terms. Pre-orders are paid at purchase and shipped within two weeks after printing.
This book may contain mistakes. You can help correct it.
Nobody is perfect, and neither is a technical book of this scale. It may contain mathematical, code, data or editorial errors. If you find one, report it publicly so it can be checked, discussed and corrected.
Report an error on GitHub Issues ↗- Review status
- Public review edition. The book is prepared in public so that readers can report mathematical, technical, code and editorial defects through GitHub issues. A public-review release is a review copy, not a claim that every chapter has completed final validation. The author reviews reported issues and updates the book as verified corrections are made.
- Rights
- Text under CC BY-NC-ND 4.0 — free to read and to share verbatim, non-commercially, with attribution. Code and tool contracts under MIT. An authorized print edition is available from the author.
If you have questions
Practical answers about the PDFs, code, licenses and printed edition.
How do I get the PDFs?
Enter your email once and we will send a link to the complete digital edition. From that page you can download the whole book or any of the fourteen chapter PDFs separately. This is a one-time delivery email, not a marketing subscription.
Is the digital version really the whole book?
Yes. The digital edition contains both volumes and all fourteen chapters. It is not a shortened preview. The printed set contains the same core book for readers who prefer a physical reference and want to support the project.
Will downloading the book subscribe me to a newsletter?
No. Your email is used to deliver access to the PDFs. QuantJourney research & updates are a separate, optional choice on the form. Leaving that checkbox empty does not add you to an updates list.
Can I share or cite the book?
The book text is available under CC BY-NC-ND 4.0. You may share unmodified copies for non-commercial purposes with attribution. You may also cite the book in research, teaching and technical work. The code has a separate MIT license.
Can I use the code at work?
Yes. The Python, C++ and SQL implementations, JSON tool contracts and validation tools are MIT-licensed, including for commercial use. The license permits use; it does not replace your own review, security, model-risk or production-approval process.
What makes a question auditable?
A question does not stop at a short interview answer. It includes the derivation, assumptions, failure modes, an implementation contract, executable code and tests where the claim can be checked numerically. The public repository lets you inspect and run that evidence yourself.
Is the code ready to put directly into production?
Treat it as reviewed educational reference code, not as a drop-in trading system. The implementations and tests make assumptions visible, but every firm must validate data, numerical behavior, controls, performance, security and operational risk in its own environment.
Was this written by AI?
AI is part of the research, drafting and engineering toolchain. That is one reason the book is built around executable evidence rather than a request for trust. Numerical claims are tied to code and tests, and public review can still find mistakes. Inspect the sources, run the suite and report anything that does not hold up.
What if I find a mistake?
Please open a GitHub issue with the chapter, question identifier and enough detail to reproduce the problem. Mathematical, code, data, citation and editorial corrections are all useful during the public review. The issue trail keeps the discussion visible and checkable.
Why two volumes?
Two durable volumes make the book practical to use at the desk. Volume I covers mathematics and markets (chapters 1–7); Volume II covers research and systems (chapters 8–14). Together they form one book, but each volume can be used independently.
When does the print edition ship?
Pre-orders are paid at purchase through Stripe. After payment, you confirm delivery details on a secure QuantJourney page. The two-volume set is shipped within two weeks after printing; delivery timing after dispatch depends on the destination and carrier.