ADEYEMI SONUBI
London,United Kingdom
Summary
Model validation manager with expertise in quantitative finance and AI/ML. Strong background in independent validation of financial models and delivery of production-grade AI solutions. Demonstrates ability to enhance model robustness and governance standards while translating complex quantitative concepts into actionable insights.
Overview
1
1
Certification
8
8
years of post-secondary education
11
11
years of professional experience
Work history
Model Validation Manager
Aviva Investors
London
2024.01 - 2026.05
- Lead independent validation of multiple model classes across investment management, including IRB, ESG, Portfolio Optimisation, Commercial Mortgage, and Pricing & Valuation models.
- Delivered 10+ end-to-end validation reviews, producing regulator-aligned reports with clear findings, limitations, and actionable recommendations.
- Identified material model risks and data issues, driving improvements in model robustness, documentation, and governance standards.
- Constructively challenged model design through quantitative benchmarking, sensitivity testing, and challenger model development, including feature simplification to improve interpretability and performance.
- Developed an AI-driven financial crime tool to improve operational efficiency and model interpretability
Associate Quantitative Finance Editor
Risk.net
London
2018.11 - 2024.01
- Managed peer review of 200+ quantitative finance research papers spanning credit risk, counterparty credit risk, market risk, derivatives pricing, structured products, and computational finance.
- Evaluated economic and financial models used by industry practitioners, assessing methodology, assumptions, and robustness.
- Maintained strong relationships with academic and industry stakeholders across global financial institutions.
- Developed deep domain knowledge of model risk management practices, regulatory requirements, and emerging quantitative methodologies.
Data Scientist – Credit Risk (Intern)
Vector ML Analytics
London
2020.09 - 2021.03
- Developed loan default prediction models using supervised machine learning (Logistic Regression, Random Forest, Gradient Boosting) in Python.
- Conducted end-to-end data preprocessing, feature engineering, model training, and validation.
- Estimated Expected Credit Loss (ECL) under IFRS 9 framework using PD, LGD, and EAD modelling.
Doctoral Researcher
University of Milan-Bicocca
Milan-Bicocca, Italy
2014.12 - 2018.03
- Optimised mixed energy technology portfolios using stochastic programming and analysed market risk via VaR, Expected Shortfall, and CVaR Deviation.
- Developed agent-based models and conducted Monte Carlo simulations for pricing and risk quantification.
- Performed statistical analysis of financial time series data using MATLAB and R.
- Published four research papers in top-tier peer-reviewed journals and presented at international conferences.
Education
PhD - Quantitative Finance
University of Milan-Bicocca
Milan-Bicocca, Italy 2014.01 - 2018.01
MSc - Mathematical Finance
Swansea University
UK 2012.09 - 2014.01
BSc - Mathematical Sciences
FUNAAB
Nigeria 2005.10 - 2008.01
Skills
Programming & Data: Python, SQL, R, MATLAB
ML & AI Frameworks: Scikit-learn, XGBoost, TensorFlow/Keras (exposure), Statsmodels
Data Engineering: Pandas, NumPy, SciPy, Feature Engineering, ETL Pipelines
Visualisation & BI: Streamlit, Plotly, Matplotlib, Seaborn, ydata-profiling
Governance & Risk: Model Risk Management (MRM), SR 11-7/SS1/23, IFRS 9, EU AI Acts
Collaboration: Git, Technical Report Writing
Certification
- Level 7 Apprenticeship in Artificial Intelligence & Data Specialist — Completed 2025
- Create Machine Learning Models in Microsoft Azure — Microsoft — Issued Aug 2023
- Data Science Analyst (DSS) — Corporate Finance Institute (CFI) — Issued Oct 2023
RECENT AI & DATA PROJECTS
Payment Alert Triage Platform – Financial Crime | ML-Powered Decision Support
- Built an end-to-end Streamlit application to automate the review, scoring, and prioritisation of payment alerts, improving speed, consistency, and decision quality for compliance teams.
- Designed a multi-model ensemble engine combining rule-based typologies (45%), Isolation Forest anomaly detection (35%), and Logistic Regression (20%) to detect both obvious and complex risk patterns.
- Embedded explainability and auditability, providing clear, regulator-ready rationale for every alert score, rule trigger, and feature contribution (SAR-ready outputs).
- Implemented a structured 5-stage workflow (Data → Risk Config → Review → Deep-Dive → Governance) aligning with operational and control processes.
- Delivered measurable efficiency gains by reducing manual review effort, accelerating alert prioritisation, and improving accuracy in identifying higher-risk transactions.
- Tech: Python, Streamlit, Scikit-learn, Pandas, NumPy
ML Framework Application (CRISP-DM) – Reusable ML Project Delivery Platform
- Designed and built a modular application to standardise and accelerate end-to-end machine learning delivery across the full CRISP-DM lifecycle.
- Implemented a 7-stage structured workflow (Business Understanding → EDA → Preprocessing → Feature Engineering → Modelling → Evaluation → Prediction), improving consistency and governance in model development.
- Embedded automated data profiling (ydata-profiling) to enhance data quality assessment and reduce manual exploratory analysis effort.
- Enabled multi-model training and benchmarking (classification, regression, clustering) with cross-validation and hyperparameter tuning, supporting robust model comparison and selection.
- Optimised performance through lazy loading, caching, and data sampling, ensuring scalable and responsive analysis for large datasets.
- Increased analyst productivity by reducing development time, improving reproducibility, and standardising best practice ML workflows.
- Tech: Python, Streamlit, Scikit-learn, Plotly, SciPy
Timeline
Model Validation Manager
Aviva Investors
2024.01 - 2026.05
Data Scientist – Credit Risk (Intern)
Vector ML Analytics
2020.09 - 2021.03
Associate Quantitative Finance Editor
Risk.net
2018.11 - 2024.01
Doctoral Researcher
University of Milan-Bicocca
2014.12 - 2018.03
PhD - Quantitative Finance
University of Milan-Bicocca
2014.01 - 2018.01
MSc - Mathematical Finance
Swansea University
2012.09 - 2014.01
BSc - Mathematical Sciences
FUNAAB
2005.10 - 2008.01