Advanced IFRS 9 Credit Risk Model Development & Validation with Python

Nitin Kumar

IFRS 9 Credit Risk Modelling with Python

Build and Validate production-grade ECL models using Python- PD, LGD & EAD

IFRS 9 fundamentally changed how banks and financial institutions recognise credit losses. Where IAS 39 let firms wait until a loss had already occurred, IFRS 9 demands a forward-looking Expected Credit Loss (ECL) model - and getting it wrong has direct P&L consequences, regulatory scrutiny, and audit risk.

Yet most practitioners learn IFRS 9 piecemeal: a seminar here, a whitepaper there, no coherent framework that takes you from raw data to a fully validated ECL engine. This course closes that gap.

In 18 hours of live, instructor-led sessions you will move through the entire IFRS 9 modelling lifecycle - from the three-stage impairment framework to PD, LGD, and EAD modelling, macroeconomic scenario integration, and regulatory-grade validation. Every concept is implemented in real code (Python) on realistic datasets, so you leave with skills you can apply on Monday morning.

Whether you are building models for the first time, validating someone else's work, or preparing for a regulatory review, this course gives you the technical depth and practical confidence to do it well.

What you’ll learn

Explain the IFRS 9 framework end-to-end -the three-stage impairment model, SICR assessment, and how it differs from both IAS 39

Build Probability of Default (PD) models using logistic regression, WoE/IV transformations, fine and coarse classing using Python

Develop Loss Given Default (LGD) models for secured and unsecured exposures, including downturn LGD, beta regression, and two-stage cure

Model Exposure at Default (EAD) across term loans, revolving facilities, and off-balance sheet items using credit conversion factor (CCF)

Incorporate forward-looking macroeconomic scenarios (base, upside, downside) into PIT-PD satellite models and calculate ECL

PD Model validation
LGD Model validation
EAD Model validation

Learn directly from Nitin

Nitin Kumar

Contact

Credit risk professional with deep expertise in IFRS 9 and credit risk modelling

See all products from Digital

Who this course is for

Credit risk analyst or modeller looking to move beyond IAS 39 and build rigorous, compliant ECL models
Model validation specialist who needs to assess PD, LGD, and EAD models against regulatory and best-practice standards
Financial controller or accountant wanting to understand the quantitative mechanics driving IFRS 9 numbers

What's included

Live sessions

Learn directly from Nitin Kumar in a real-time, interactive format.

Certificate of completion

Share your new skills with your employer or on LinkedIn.

Maven Guarantee

Your purchase is backed by the Maven Guarantee.

Course syllabus

Week 1

Sep 12—Sep 13

Session 1 — IFRS 9 Fundamentals

1 item

Session 2 — IFRS 9 Data Preparations and PIT PD models

1 item

Week 2

Sep 14—Sep 20

Session 3 — Loss Given Default (LGD) Modelling

1 item

Session 4 — EAD Modelling

1 item

Schedule

Live sessions

18 hrs / week

Build production-grade ECL models from scratch — PD, LGD, EAD, scenario analysis, and validation — in 6 intensive live sessions with hands-on Python practice

Projects

36 hrs / week

- Build Probability of Default (PD) models using logistic regression, WoE/IV transformations, fine and coarse classing, and survival analysis — in Python - Develop Loss Given Default (LGD) models for secured and unsecured exposures, including downturn LGD, beta regression, and two-stage cure/recovery models

Async content

12 hrs / week

Incorporate forward-looking macroeconomic scenarios (base, upside, downside) into PIT-PD satellite models and calculate probability-weighted ECL

Testimonials

I'd been producing IFRS 9 numbers for two years without truly understanding where they came from. This course gave me the language, the logic, and the code to own the whole model lifecycle.
Sarath Kumar
Senior data scientist
I could tick boxes on a validation template, but I couldn't look a model developer in the eye and say 'your downturn LGD assumption is wrong, and here's why.' Now I can — and I do.
Prashanth Tiwari
Senior Credit risk modeller
My auditors were asking questions I was passing straight to the quant team. After this course, I could answer them myself — and challenge the model outputs when they didn't make commercial sense.
Anshuman Misra
Senior Model validation manager