Skip to content
JoaFin
ES/EN

The firm

A quantitative practice, not a reporting service

JoaFin works at the boundary between mathematics and the decision: the same problems go through the blackboard, the trading desk and production before they count as solved.

JoaFin is a quantitative finance practice based in Mexico City. We work with trading desks, risk functions, insurers, power market participants and law firms, on projects where the result has to hold up in front of someone: a committee, an auditor, a regulator or a court.

The foundation is mathematical, and not decorative. The practice was built on doctoral research in mathematical analysis and optimal transport, with two published papers, and that origin sets the standard: a method is adopted for what can be proven about it, and a number ships only when it can be reproduced from the raw data.

The team pairs that training with years spent where rigor gets paid or punished quickly. An energy desk, on two fronts at once: locational marginal price forecasting on one side and daily statistical arbitrage on the other, with risk limits set from the tail properties of the outcome distribution. Equities, where research has to become a position with money behind it. And senior quant work at a derivatives analytics platform, where local volatility calibration is solved through optimal transport: arbitrage-free surfaces, calibrated in real time, feeding pricing, volatility forecasting and Value-at-Risk.

As a consultancy we have built locational marginal price and electricity demand models in Mexico, actuarial valuations, and quantitative analysis presented before a court. Alongside that we build the infrastructure those models need: scrapers, pipelines, databases and complete applications. We would rather hand over something that works than a document with recommendations.

We use AI intensively across the whole work cycle, and it has changed a great deal about what a small team can deliver. What has not changed is where confidence in a result comes from: the mathematics holding it up, and the verification backing it.

Why JoaFin exists

Because a lot of people are using AI badly. They use it to skip past framing the problem instead of reaching a well-framed one sooner, and then cannot say why their model ought to work at all. That produces results which look fine right up until the day they matter.

JoaFin is the quantitative consulting that always worked — models, risk, valuation, infrastructure — amplified by these tools rather than replaced by them. The speed gain is real and we take it. What does not change is that somebody answers for the result, by name and with an argument.

Experience on the team

Where each practice area comes from. Not a list of interests: work done in the seat, with money or a signature on the line.

  1. Derivatives

    Senior quant · Derivatives analytics platform

    Core team. Optimal-transport-based local volatility calibration: arbitrage-free, real-time, integrated into pricing, volatility forecasting and VaR.

  2. Energy markets

    Power market trading and analysis

    Locational marginal price forecasting and daily statistical arbitrage, with risk limits set by the tail properties of the outcome distribution. Price formation, network congestion, hedging and portfolio risk.

  3. Equities

    Research and trading

    Fundamental and quantitative research translated into position and execution decisions.

  4. Academia

    Research and teaching

    Optimal transport, mathematical analysis and its applications. Publications in PDE analysis and in machine learning; undergraduate and graduate teaching.

  5. Consulting

    JoaFin projects

    Locational marginal price and electricity demand models, actuarial valuation, expert quantitative analysis, data platforms and application development.

Teaching

Part of the team has taught undergraduate and graduate courses. The range matters, and not as decoration: explaining complex analysis and general relativity well forces genuine understanding — and it is the same muscle needed to explain a model to a committee that is not technical.

  • Complex analysis
  • Linear algebra
  • Calculus
  • Numerical optimization
  • General relativity
  • Optimal transport (graduate)

Toolkit

Statistics
Inference, time series, survival analysis, extreme values and heavy tails, validation design, uncertainty quantification, actuarial mathematics.
Machine learning
Predictive modeling and model selection, Γ-convergence of parametric models, domain adaptation and transfer learning, WGANs, drift monitoring. And language models as a daily implementation tool.
Mathematical finance
Local volatility and arbitrage-free surfaces, derivatives pricing, VaR and CVaR, stochastic analysis, portfolio risk, statistical arbitrage.
Mathematical methods
Constrained and unconstrained optimization, PDEs and numerical methods, probability, and optimal transport: Monge–Kantorovich, Wasserstein metrics, entropic regularization, gradient flows.
Languages
Python, SQL, TypeScript/JavaScript, R, MATLAB, C++.
Data and backend
PostgreSQL, ETL pipelines and orchestration, scraping, APIs, cloud deployment, Next.js and React, dashboards and reproducible reporting.

Have a problem that spreadsheets won't settle?

Get in touch