About
The gap between knowing the theory and working the desk is wider than it looks.
Brownian Bridge is a quantitative finance platform built to close it — through working implementations, rigorous derivations, practical projects, and observable skill signals that a quant hiring team would recognise as credible.
The gap
Most quant programmes produce graduates who can cite the models. Few can implement them under real conditions.
The problem is not intelligence or effort. Academic programmes are optimised for mathematical rigour, not production-grade practice. There is nothing wrong with studying Girsanov in the abstract. But the candidates who succeed on a quant desk are also the ones who can do what follows.
What most programmes cover well
- Measure theory and Itô's lemma
- Black-Scholes derivation from first principles
- Heston, SABR, and rough volatility theory
- PDE methods and transform techniques
- Monte Carlo convergence proofs
What desks actually evaluate
- Calibrate a vol surface under time pressure
- Debug a finite-difference solver at the boundary
- Explain a PnL attribution discrepancy
- Write pricing code that survives model review
- Discuss model limitations in a live interview
What we do differently
Not a course catalogue. A production environment for quant skill.
Traditional certifications reward completion. Brownian Bridge rewards demonstrated ability. The code you write, the models you implement, the challenges you solve — all of it is inspectable. That is the point.
Working code, not pseudocode
Every implementation compiles, runs, and ships with unit tests. C++20 and typed Python — production-grade by default.
Derivations, not definitions
Theory is derived, not stated. Assumptions are explicit before any formula appears. Every equation earns its place.
Reproducible notebooks
Every numerical result runs end-to-end in a clean kernel with a fixed seed. No black-box outputs.
Observable skill
Work product — code, derivations, challenge solutions — is inspectable. Skill is demonstrated, not self-reported.
Peer-reviewed contributions
Content is held to the standard of a quant research note. Depth and correctness are rewarded. Volume is not.
Interview-calibrated content
Every topic includes how it appears in real quant interviews — at junior, senior, and quant researcher level.
Content standards
Every piece of content on the platform meets the same bar.
These are not aspirational guidelines. They are enforced requirements. Content that fails them is not published.
Assumptions stated before derivation.
No model appears without its mathematical and market context made explicit.
Code that compiles and runs.
With unit tests. Pseudocode is labelled as such.
Limitations and failure modes discussed.
A model without known failure modes described is incomplete content.
Numerical results reproducible.
With stated seeds, tolerances, and environment specifications.
Conventions always made explicit.
Compounding basis, day count, volatility units — specified, not assumed.
For hiring teams
Skill you can inspect. Not credentials you have to trust.
Quant hiring is expensive and slow because the standard signals — CVs, university names, certifications — are poor proxies for what a candidate can actually do on a desk. Brownian Bridge is building a different signal layer.
Candidate skill data on this platform is derived from observable behaviour: problems solved, code submitted, peer reviews conducted, challenge rankings achieved. Hiring teams can inspect the work directly — the implementation, the derivation, the test suite — not a summary of it.
Brownian Bridge does not certify candidates. It makes their technical output visible to teams who know what to look for.
Who built it
Benoit Vandevelde
Director, Quantitative Modelling — Dymon Asia Capital
Brownian Bridge was built by someone who has spent more than a decade on both sides of the gap it targets. During the week: building and validating pricing models at Deutsche Bank, Barclays, BNP Paribas, ICBC Standard Bank, and Dymon Asia Capital. In term time: lecturing financial mathematics and C++ to Master's students at Université Gustave Eiffel in Paris — implementing models by Heston, Dupire, Andersen, and others in production-grade code.
Having assessed candidates from both sides — as a practitioner who has interviewed and as an educator who has seen what students can and cannot do — the gap between academic training and desk-ready skill becomes impossible to ignore. The platform exists because closing that gap requires the practitioner's standard and the teacher's method in the same place.
Career
Mar 2025 – present
Director, Quantitative Modelling
Dymon Asia Capital, Singapore
Sept 2023 – Mar 2025
Director, Cross-Asset Model Validation
ICBC Standard Bank, London
Sept 2019 – Aug 2023
Team Lead, Commodities Model Validation
Deutsche Bank, London
May 2017 – Aug 2019
VP, Front Office Quantitative Analyst, CCR
Barclays, London
Jul 2016 – Apr 2017
Quantitative Analyst, R&D
Numerix, London
2014 – 2016
Quantitative Analyst, Pricing & Model Validation
Lloyds Bank, London
2013 – 2014
Quantitative Analyst, Front Office
BNP Paribas, London
Jan 2016 – present
Lecturer, Financial Mathematics & C++ for Derivatives Pricing
Université Gustave Eiffel, Paris — Master's level (parallel to industry roles)
Education
Master 2, Mathematics and Applications in Finance — Ecole Nationale des Ponts et Chaussées
Engineering diploma (Finance & CS majors) — Ecole Centrale de Nantes
Where this is going
A reference point for quantitative skill — built on work product, not credentials.
The long-term goal is not a course catalogue. It is a living quantitative ecosystem: a platform where quant students build verified, production-grade portfolios of work; where experienced quants contribute, review, and build external reputation through depth of insight; and where employers access skill data derived from observable behaviour rather than self-reported competencies.
The target state is a platform that becomes the default pre-screening step for quant roles — not because it markets itself as such, but because the work product it surfaces is genuinely interview-grade and desk-ready. That standard is not an aspiration. It is the admission requirement.
Build evidence of what you can do.
Courses, coding challenges, interactive labs, and notebooks — all held to the same standard a quant desk would apply.
Brownian Bridge — independent platform, founded 2025.