📚 New · Public LMS

Apprenez les maths. Reproduisez les papiers. Tradez avec le même Rust qu'un desk institutionnel.

Six parcours, du cours zéro à la stratégie quant — chaque leçon compile en Rust et tourne dans le moteur que nous utilisons en production.

Learn the math. Reproduce the papers. Trade with the same Rust stack as an institutional desk.

Six learning paths, from zero to PhD-level quant — every lesson compiles to Rust and runs inside the same engine we ship in production.

Neuf parcours, une stack

Nine paths, one stack

Tous les parcours utilisent le même hfthot_lab_core Rust et le même moteur d'exécution. Pas de jouet : ce que vous apprenez, vous le faites tourner en live.

Every path runs on the same hfthot_lab_core Rust crate and the same execution engine. No toys — what you learn, you ship.

📈
Trader+

Microstructure & OU

Solve a 12×12 matrix Riccati in Rust, in milliseconds, on live LOB.

2 leçons · OU prerequisites · Cross-asset blog
🎯
Researcher+

MFG & Optimal Control

From DE to HJB to MFG — full pipeline on synthetic flows.

3 leçons · Differential evolution · HJB · Mean-field games
🌊
Researcher+

Rough Vol & Signatures

Rough Heston, signature kernels, fast iteration on synthetic paths.

3 leçons · Hurst exponent · Path signatures · Kernel methods
🧩
Pro+

Topology, gentle intro

Persistent homology on a toy basket. Theory, math, intuition. Calibrated WP6 stays private.

1 leçon · Theory only · Synthetic basket
🛠️
Trader+

Engineering for quants

Polars + Rust + Criterion. Beat the published benchmarks.

3 leçons · Polars patterns · Rust kernels · Criterion benches
🧪
Trader+

Practice — paper trading

Pick your broker. Keep your data. Run the engine.

2 leçons · Multi-broker sandbox · Reproducible notebooks
📐
Trader+ FREE

Rappel maths — Martingales & vol rugueuse Maths refresher — Martingales & rough vol

Espaces de probabilité, calcul d'Itô, mouvement brownien fractionnaire. Bilingue 🇫🇷/🇬🇧.

Probability spaces, Itô calculus, fractional Brownian motion. Bilingual 🇫🇷/🇬🇧.

2 lectures · Prérequis pour Rough Heston2 readings · Prerequisite for Rough Heston
⚛️
Pro+ CHF 49

Physique théorique — De Dirac à la gravité quantique Theoretical physics — From Dirac to quantum gravity

Spineurs, équation de Dirac, Wheeler-DeWitt, supergravité, TQFT — et leur lien avec la finance quantitative. Bilingue 🇫🇷/🇬🇧.

Spinors, Dirac equation, Wheeler-DeWitt, supergravity, TQFT — and their link to quantitative finance. Bilingual 🇫🇷/🇬🇧.

8 chapitres · Notebook d'exercices · Achat unique CHF 498 chapters · Exercise notebook · One-shot CHF 49
🌊
Researcher+ CHF 199

Rough Heston Premium — Arbitrage de volatilité Rough Heston Premium — Volatility arbitrage

Noyaux de Volterra, Riccati fractionnaire, pricing FFT, calibration sur surface IV publique S&P 500. Inclus dans le tier Researcher.

Volterra kernels, fractional Riccati, FFT pricing, calibration on the public S&P 500 IV surface. Included with the Researcher tier.

2 leçons · Paper premium PDF · Notebook compagnon · CHF 199 ou inclus dans Researcher (CHF 15/mois)2 lessons · Premium paper PDF · Companion notebook · CHF 199 or included with Researcher (CHF 15/mo)

Notre promesse de recherche

Our research promise

Pas de "magic alpha" caché. Pas de courbe en arrière. Tout est reproductible.

No hidden "magic alpha". No back-fitted curves. Everything is reproducible.

🔬 Public datasets & synthetic only

Every LMS lesson uses public market data or synthetic processes. Calibrated proprietary models stay in the private instance build — see Ethics.

🦀 Same Rust as production

The crate hfthot_lab_core that powers the LMS is the exact same crate that drives our HFT execution. Verify the SHA on GitHub.

📓 Fork as notebook

Every lesson and every paper-trading session can be forked as a Jupyter notebook in one click. Your iterations stay yours.

🏗️ Compose with your data

Polarway gives you a typed parquet+gRPC client. Bring your own lakehouse, plug it in, and run the LMS labs against your own flows.

A taste — Almgren-Chriss optimal execution (lesson 02)

𝓛(x, ẋ, t) = ½ λ x² + ½ η ẋ²  ⟶  ẍ* = (λ/η)¹ᐟ² x

A taste — HJB for MFG (researcher path, lesson 03)

∂ₜu + H(x, ∇u, m) − ½ σ² Δu = 0,  ∂ₜm − ∇·(m ∇H) − ½ σ² Δm = 0