Traffic and edge layer
How a quant product handles real traffic before it reaches application code.
DNS/TLSload balancersAPI gatewaysCDN cache keyssigned media URLs
MatheLinuxQuant learning system
Roadmap
Math, C++ / HPC, systems, microstructure, risk.
Advanced system design layer
How a quant product handles real traffic before it reaches application code.
Where Redis helps, where it hurts, and how to keep cached state consistent.
The event-driven layer behind payments, grading, notifications, research jobs, and data pipelines.
Production architecture for ticks, bars, features, backtests, and live services.
The service boundary between strategies, order routing, pre-trade checks, and audit logs.
Operating the platform: observing failure, deploying safely, and recovering cleanly.
The calculus a quant uses every day — limits, derivatives, Taylor series, integration, multivariable gradients and Lagrange multipliers, and vector calculus — built with full derivations and worked examples.
The other half of quant math — vectors, matrices, eigenvalues, SVD, PCA, least-squares regression, and convex optimization — built to where you can extract PCA factors and solve a Markowitz portfolio by hand.
The math a working quant actually uses — probability, differential equations, stochastic calculus, the Black–Scholes model, and numerical methods — built from the ground up with full derivations, worked examples, and graded problems.
The implementation half of quant work — modern C++, numerically-sound pricers, and the high-performance computing (cache, SIMD, threads, GPU) that production pricing engines run on.
Production architecture for quant platforms: load balancers, CDN, Redis, queues, market-data replay, execution risk, and observability.
Self-exciting events, volatility clustering, and order-book dynamics.
Regression through deep learning and transformers, applied to alpha discovery, credit risk, and backtest overfitting.
Build portfolios from expected returns, covariance estimates, constraints, transaction costs, and audit-ready risk evidence.