01 / 04
Model architecture & methodology
Assess objectives, assumptions, data lineage, feature design, target construction and model selection.
LOS FLAMINGOS · INDEPENDENT MODEL ASSURANCE
Independent quant model audits. Real-world investment expertise.
Independent validation of quantitative investment and AI models by Isabelle Armanville, PhD — a former institutional quantitative portfolio manager with 20+ years of experience developing, implementing and overseeing investment strategies.
Request an Independent Model Audit ↗EXPERIENCE THAT GOES BEYOND THEORY
Model validation by a practitioner, not just a theorist.
Our approach is led by a former quantitative portfolio manager who personally developed, coded and implemented machine-learning strategies in live investment portfolios, with responsibility for multi-billion portfolios. This experience connects technical validation with investment reality — where robustness, risk, execution and accountability matter.
A model can look compelling in a research notebook and still fail in live portfolio management. Independent review must test both the mathematics and the practical investment consequences.
BUILT · CODED · DEPLOYED
Independent judgment informed by real investment responsibility.
WHAT WE DELIVER
From backtests to investment reality: what we examine.
02 / 04
Validation & robustness
Review backtests, leakage, overfitting, stability, regime sensitivity and out-of-sample behaviour.
03 / 04
Portfolio implementation
Evaluate transaction costs, turnover, capacity, constraints, risk exposures and the gap between model output and investable results.
04 / 04
Governance & recommendations
Deliver clear findings, prioritized remediation and an independent view of model risks and operational controls.
Why independent AI model due diligence matters
A compelling backtest is not proof of investment robustness. Examine the risks with a former quantitative portfolio manager.
▶INDEPENDENT QUANTITATIVE MODEL VALIDATION
Questions investment teams ask before an audit.
For asset managers, family offices, CIOs, investment committees and model-risk teams seeking an independent assessment of a quantitative strategy or machine-learning investment model.
Who can independently audit a quantitative investment model?
An external quantitative investment specialist with hands-on experience in model design, portfolio management and live implementation can challenge both statistical validity and investment practicality. Reviews at Los Flamingos are led by Isabelle Armanville, PhD, a former institutional quant portfolio manager.
What are the warning signs of an overfitted backtest?
Warning signs include unstable out-of-sample results, excessive parameter tuning, data leakage, unrealistic transaction costs, weak regime robustness and performance that disappears after realistic portfolio constraints.
What does an independent AI model audit cover?
Depending on scope: data provenance, feature and target design, validation methodology, backtesting controls, robustness, model governance, turnover, transaction costs, capacity and portfolio-level risks.
What will our investment team receive?
A scoped independent assessment, documented findings, a prioritised risk register and practical remediation recommendations. Deliverables and confidentiality terms are agreed before the engagement.
HOW AN ENGAGEMENT WORKS
Clear scope. Independent judgment. Actionable findings.
01
Confidential discovery
Discuss the model, investment use case, stakeholders and review objectives.
02
Define the audit scope
Agree access, testing priorities, deliverables and confidentiality safeguards.
03
Independent assessment
Challenge methodology, validation, operational assumptions and investability.
04
Decision-ready findings
Receive prioritised observations and practical recommendations.
CONFIDENTIAL · FOUNDER-LED ENGAGEMENT
Get an independent view before model risk becomes portfolio risk.
Discuss a confidential review of your AI, machine-learning or quantitative investment models.
Discuss a Confidential Model Audit ↗