Skip to main content

Skills application

Instruction and application
In Progress

The performance lab​

In this skills application, you will fine-tune, ensemble, and calibrate a credit risk model to enhance its accuracy, robustness, and reliability in real-world lending decisions.

<g></g><defs><clipPath><rect width="24" height="24" fill="white"></rect></clipPath></defs>## Context Organization

You have joined SafeFinance, a digital lending company that leverages machine learning to assess credit risk and predict loan defaults.Current challenge

SafeFinance’s credit risk model achieves strong accuracy but consistently overpredicts default risk, causing the company to reject creditworthy applicants and lose valuable business opportunities. Preliminary analysis suggests that while the model is accurate on average, its probability estimates are poorly calibrated, and its performance could be more robust across different applicant segments.

Your role​

As an ML Engineer, your task is to refine and strengthen the model’s performance using advanced training strategies. You’ll applyhyperparameter tuning,ensemble methods, andcalibration techniques to balance accuracy, robustness, and reliability—ensuring the model supports confident, fair lending decisions.

Objective​

By the end of this activity, you will have enhanced the model’s predictive capability and probability reliability, demonstrating how tuning, ensembling, and calibration together drive more trustworthy and business-aligned ML outcomes.

Success criteria​

To complete this skills application successfully, you must:

  • Apply and compare hyperparameter tuning methods to identify the best-performing configuration for the model.
  • Implement an ensemble approach and evaluate how it improves accuracy and robustness over a single model.
  • Apply and assess model calibration techniques to enhance probability reliability and decision confidence.
  • Analyse trade-offs between performance, interpretability, and computational efficiency.
  • Reflect on how tuning, ensembling, and calibration together support SafeFinance’s goal of making fair, reliable lending decisions

Instructions and materials​

Follow the steps below to complete this skills application. Completing this activity will “unlock” the solution example on the following page.

Download the activity resources​

Select the links below to download the Jupyter Notebook anddataset for this activity. The notebook provides the step-by-step guidance, while the dataset file is required to complete the model tuning and calibration tasks.

Skillable_lab_M7_Unit2_Performance_Lab.ipynbsafefinance_credit_risk_dataset.csv

Access the Jupyter Notebook and dataset​

  • Open the provided Jupyter Notebook for this activity and download the accompanying dataset.
  • Upload the dataset to your notebook environment before running the code.
  • The notebook contains all baseline code and step-by-step guidance to help you complete the model tuning and calibration tasks.

Complete the tasks​

Follow the instructions in the notebook to:

  • Tune model parameters to improve performance.
  • Apply ensemble methods to enhance robustness.
  • Calibrate predictions for greater reliability.

Record and submit​

Document your findings and reflections directly in the notebook. Save your completed file and submit it for review.