Artificial Intelligence & GenAI · Machine learning

Risk Scoring ML System

Develop a cost-sensitive classifier with threshold tuning and explainability.

Capstone in Machine Learning Engineering.

Recommended effort24 hoursCourse levelIntermediateSuggested teamIndividual or team of 2-3

Get project curriculum →Review complete project scope

The project brief

The business challenge

Classification errors can have different costs. The model needs a documented threshold and explanations that show how its scores should be interpreted.

Shared method: define the requirement, design and build the solution, then test and document the result.

Project brief

Test scenarios

Use these scenarios to plan the project review. They describe intended checks, not completed learner results.

01 · Test scenario

Classification costs

Compare model results under the documented costs of false positives and false negatives.

02 · Test scenario

Threshold trade-off

Evaluate alternative score thresholds and explain the effect on classification errors.

03 · Test scenario

Score explanation

Inspect a model score with SHAP and compare the API output with the evaluated model.

Inside the working solution

What you will build

01Cost-sensitive classifier
02Threshold tuning and model explanations
03Risk-scoring API and deployment architecture

Your final submission

What you will present

Experiment report
SHAP analysis
API
Deployment architecture

Technology workspace

The tools behind the build

Review the tools and skills used in this project brief.

Python

Data processing, model logic and backend automation

scikit-learn

Use scikit-learn within guided implementation, testing and portfolio workflows

SHAP

Use SHAP within guided implementation, testing and portfolio workflows

FastAPI

Expose project logic through production-style APIs

Project brief

Acceptance criteria

Review the cost trade-offs, score explanations and deployment design.

The experiment report documents classification costs and threshold selection.
SHAP analysis explains the model results and records limitations.
The API and deployment architecture correspond to the evaluated classifier.

Questions about this project

Risk Scoring ML System FAQs

Check the recommended level, tools and guidance before selecting a programme.

Who is the Risk Scoring ML System project suitable for?+

The associated course is taught at intermediate level. Review its prerequisites before choosing this capstone. An adviser can help you confirm the appropriate starting point.

Which tools and skills are used?+

The project brief uses Python, scikit-learn, SHAP, FastAPI. Confirm the selected stack with admissions when choosing a programme.

Is mentor guidance included?+

Ask admissions to confirm the mentor reviews, feedback and final walkthrough included in your selected programme.

Request the exact training scope

Explore this project with admissions

Ask about the curriculum, prerequisites, mentor reviews and recommended programme for Risk Scoring ML System.

Project module and tool breakdown
Recommended prerequisites and programme
Demo class and upcoming batch guidance

Project curriculum request

Explore this project in training

Receive the project scope, tools, mentor review process and recommended programme for Risk Scoring ML System.

Secure enquiry. Your details are used only for admissions guidance.