Modelling / Forecasting Senior Analyst (Quants)

TD•Toronto, ON

10d•Onsite

About The Position

The Model Validation (MV) group is a centralized model risk management function within the Bank. The MV team is responsible for the vetting and approval of complex mathematical and statistical models used in credit lending, business operations, and stress testing. By ensuring an objective and independent evaluation of models, the model validation function is critical to the effective measurement and management of risk across the TD Bank Group.

Requirements

Strong statistical background and excellent analytical and problem-solving skills with a graduate degree in one or more of the following areas: statistics, economics, finance, mathematics, computer science, or engineering.
Experience with predictive model development, credit scoring, and/or AI/ML modeling
Hands-on experience with Python for statistical analysis and modeling or SAS is a must (Python preferred)
Excellent verbal and written communication skills
Good time management and multitasking skills
Quick learner who grasps new concepts and techniques quickly
Must be a good team player
Graduate degree in Statistics, Actuarial Sciences, Econometrics, Computer Science, or other quantitative fields.
2+ years relevant experience.

Nice To Haves

Familiarity with retail banking products, customer behaviors, and credit risk management is a definite asset
Experience with liquidity forecasting and liquidity stress testing models is a major plus
Familiarity with AI productivity tools such as Microsoft Copilot and GitHub Copilot is a plus

Responsibilities

Perform validation of all models deemed in-scope by the bank-wide Model Risk Policy. These models are used in the Bank for a variety of purposes, including scoring credit risk (acquisitions and account management) and Economic and Regulatory Capital (e.g. Probability of Default, Exposure at Default, Loss Given Default), conducting stress testing, detecting fraud behaviors, marketing retail products, and projecting Pre-Provision Net Revenue (PPNR). These models may use traditional statistical methodologies and/or AI/ML approaches.
Develop independent benchmarks for use in the validation of the models listed above. These benchmark models could be traditional supervised learning methods, unsupervised learning, or machine learning algorithms.
Assess the appropriateness of the model for its specific use, reasonableness of the model assumptions and the accuracy of the model implementation.
Prepare detailed reports describing the mathematical underpinnings of the model, validation techniques employed, test results obtained, and any model limitations noted.
Prepare management summaries highlighting the outcome of the validation process for each model and outlining recommendations for approval or further improvements.
Establish and maintain productive working relations with internal model development groups and external vendors who have developed customized models for TD.
Play a key role in ensuring the appropriate use of risk models.
Identify the need to implement new models/techniques for risk management as industry standards evolve and regulatory requirements change.
Stay current in knowledge of credit risk management methodologies, predictive modeling, and statistical analysis.