Quantitative Economic & Financial Consulting

Models you can defend. Decisions you can explain.

Hi — I'm Brett. Economist by training, physicist by first degree, and for nearly fifteen years I've built the models that price credit, volatility, and regulatory capital for some of the world's largest asset owners, insurers, and fintech lenders.

I'm happiest in the weeds — and I take on a small number of private consulting engagements for clients whose problems live there too: bespoke modeling, quantitative research audits, and decision support under uncertainty.

Credit · Quant · Risk · Triathlon
Start a conversation or see what I do ↓
01 Services

What can I do for you?

Every engagement is bespoke — scoped around the decision you actually have to make, not a productized template. Four areas cover most of the work; the common thread is rigor you can put in front of a board, a regulator, or your own risk committee.

Economic & financial modeling

Need a model built — or an inherited one you no longer quite trust?

Derivatives and volatility — risk-neutral surface calibration at scale, realized-measure modeling — credit models capturing the interaction of default, loss, spreads, and interest rates, and macroeconomic scenario generation.

I directed the models that became the insurance sector's benchmark for negative interest rates — a $10M+ revenue stream — and I've since calibrated option-implied surfaces minute-by-minute across an entire equity universe. You get the model, the calibration machinery, and documentation your team can own. Not a black box.

Quantitative research audits

Would your backtest survive contact with reality?

Independent review of models, backtests, and research pipelines: experiment design, pre-registered predictions, honest cost and slippage measurement, leakage and artifact detection.

In my own research I've refuted far more strategies than I've deployed — that is rather the point. A result that survives an adversarial audit is one you can size with confidence; one that doesn't is a loss you didn't take. I'll tell you which you have — plainly, and with the evidence.

Decision support under uncertainty

The analysis is done — so what should you actually do?

Translating models into decisions: risk frameworks, capital and regulatory impact — including statutory analysis for insurers and reinsurers as the National Association of Insurance Commissioners (NAIC) evolves its risk-based capital (RBC) framework — and briefings a board can act on.

My habit, on the record: I'll give you the strongest version of both sides of an argument, and then I'll commit to a view. Even-handed is not the same as neutral.

Cloud-scale computation

What happens when the model outgrows the laptop?

AWS fleet orchestration, large-scale data pipelines, and GPU-accelerated calibration and simulation. I've cut model calculation times by 75% with Spark and CUDA, and these days I run calibration fleets over terabyte-scale tick archives — engineered for cost as carefully as for speed, because the compute bill is part of the model too.

02 Writing & speaking

Recent public work

The private engagements stay private — but some of the work speaks for itself. At Bridgeway I also wrote a weekly Newsreel and the deep-dive column In the Weeds with Brett Manning, covering statutory accounting, RBC, and IFRS 9/17 for insurance asset managers.

Plaid · December 2025

How we built LendScore: turning cash flow data into credit risk insights

The build behind a lender-ready credit risk score — 145 engineered attributes over 1.44 million tradelines, with SHAP-based explainability and fairness safeguards designed in from the start.

Read the post → Plaid · April 2026

Meet the new engine behind Plaid Income

Co-authored the announcement of the transformer-based engine that reads bank transaction data to verify income — a step-change in accuracy for income categorization.

Read the post → Plaid Effects · 2026 · Panel

How the experts get model development right

Hosted an AI risk management panel with leaders from FICO, Affirm, and 2OS — on managing bias in model development and building in feedback loops and human review without slowing shipping down.

Watch the session →

More in print

On the speaking side: five sessions on risk analytics and modeling for the Society of Actuaries. Collaborations, talks, and lectures are all welcome.

03 Career

Where has the work taken me?

The condensed version — each role expands if you want the detail.

Jul 2025 – Present Plaid — Product Manager, Credit Building the next generation of credit insights from cash flow data. +
  • Led the build of LendScore, Plaid's credit risk score — 145 engineered attributes over 1.44 million tradelines, from raw bank transactions to a compliant, explainable model in lenders' hands.
  • SHAP-based adverse-action explainability and fairness safeguards designed in from the start, not bolted on.
  • Co-authored the launch of the transformer-based engine behind Plaid Income; hosted the Effects 2026 panel on AI risk management.
Nov 2022 – Present Bridgeway Analytics — Predictive Analytics Principal, now Technical Advisor Helping insurers and asset managers navigate insurance asset regulation. +
  • Quantified the impact of new regulatory treatment on insurer holdings of CMBS, RMBS, CLOs, and other asset-backed securities — published across three industry publications.
  • Wrote a weekly Newsreel on the evolving regulatory landscape — statutory accounting, RBC, IFRS 9/17, and US GAAP.
  • Built an AI asset-categorization tool (Python on AWS, SQL backend) using NLP to identify the appropriate NAIC filing schedule; two clients engaged on proof-of-concept.
  • Connected the CMS, payments, hosting, and email systems into one automated workflow via REST APIs.
  • Since December 2023: part-time technical advisor, focused on European and Bermudan rules alongside the quantitative impact of US changes.
Dec 2023 – Jul 2025 Upstart — Principal Product Manager, Machine Learning The bridge between ML underwriting and the broader business. +
  • Translated model advances into business strategy — and business requirements back into modeling strategy — for a leading AI lending platform.
  • Drove a comprehensive data acquisition strategy, including vendor management and negotiations on 7-figure contracts.
  • Led strategic initiatives on the core pricing engine and raising the maximum amount per loan.
Apr 2022 – Feb 2023 GIANT Protocol — Lead Product Manager, Tokenomics Modeling a liquid marketplace for the $2 trillion bandwidth market. +
  • Defined and executed the research roadmap for Monte Carlo models (Python) — presented to investors for the next funding round, and closed ten protocol attack vectors along the way.
  • Improved conversion on targeted marketing by 75% using machine learning.
  • Translated the whitepaper into technical specifications for a Layer-1 Substrate implementation.
  • Advised on scaling the opportunity tenfold by building a derivative market for bandwidth on the token contracts.
2019 – 2022 Moody's Analytics — Director Research & strategy lead for a new issuer-level credit portfolio tool. +
  • Created a new seven-figure revenue stream — negotiating buy-in across lines of business and access to intellectual property from five legacy products.
  • Built and commercialized a beta simulation tool (C#/C++); sold to three new clients.
  • Developed and calibrated models capturing the interaction of credit (probability of default and loss given default), spreads, and interest rates — pricing vanilla, callable, and structured securities with industry diversification.
  • Cut calculation times by 75% by introducing Spark and CUDA to model estimation and simulation.
  • Presented five sessions on risk analytics and modeling for the Society of Actuaries; as Pride BRG co-chair, secured enhanced and equitable surrogacy benefits for 5,000+ US-based staff.
2012 – 2019 Moody's Analytics — Assistant, then Associate Director Technical lead for model calibration — Edinburgh, then San Francisco. +
  • Technical lead for the 'Own Views' team — two quants and four developers automating the calibration of risk, finance, and pricing models to each client's capital market assumptions.
  • Developed multi-start, stochastic, and uphill-step enhancements to an in-house Levenberg–Marquardt optimizer — more than halving analyst interventions during calibration.
  • Built ML asset-allocation tools — neural nets and random forests constructing efficient risk-return frontiers to maximize return on equity at a given Value-at-Risk or economic capital.
  • Spotted early that negative interest rates would break the core models — then directed the multi-disciplinary team that established the sector's benchmark models (displaced multi-factor Black-Karasinski, LMM, and CIR), securing a $10M+ revenue stream.
  • Defined and coded a VAR/ARIMA macroeconomic framework for the Economic Scenario Generator — five clients, $2M in annual revenue.
  • Company-wide Create Confidence award — six winners among 10,000 staff.
2006 – 2012 Before all that — Durham University & the classroom A PhD on inequality and financial crises — funded by six years of teaching. +
  • PhD in Economics at Durham, asking a simple question with an un-simple answer: how does economic inequality affect the stability of the financial system?
  • Delivered seminars and lectures in maths, statistics, macroeconomics, international finance, and econometrics — designing the computer practicals for econometrics at BSc and MSc level.
  • Six years of freelance teaching from GCSE to MSc — rated outstanding in an Ofsted inspection — while learning the basics of running a business: accounting, marketing, and regulation.

Education: PhD in Economics and MSc in Economics & Finance, Durham University · BSc in Physics with Theoretical Physics, Imperial College London.

"God does not care about our mathematical difficulties. He integrates empirically."
Albert Einstein

So — what's next?

I keep the practice deliberately small: a few engagements at a time, chosen for fit. If you have a modeling problem, a research program that needs an independent set of eyes, or a decision that deserves better analysis than it's getting — tell me about it. Worst case, you'll get an honest opinion.

And if you'd rather talk Ironman splits than interest-rate models, that's welcome too.