Artificial Intelligence
BurnoutSense — Developer Burnout Risk Classifier
BurnoutSense is a production ML classifier trained on 7,000 real developer profiles to detect burnout risk before it becomes critical. A Random Forest pipeline with KNN imputation processes 11 signals — stress level, screen time, sleep, caffeine, commits, meetings, and more — and returns a labelled risk class with full probability breakdown across all three categories. Stress level alone accounts for 76% of feature importance. Try it live below.
ML ModelsClassificationHealth TechDeveloper Tools
2025
2 months
Open Source
Key Features
- 3-Class Output: Predicts High, Medium, or Low burnout risk with per-class probabilities
- 99.59% Accuracy: Random Forest outperformed Gradient Boosting and Logistic Regression in grid search
- 11 Input Signals: Age, experience, work hours, sleep, caffeine, bugs, commits, meetings, screen time, exercise, stress
- KNN Imputation: Handles missing feature values with k=5 nearest-neighbour imputation
- Dominant Signal: Stress level alone contributes 76% of total feature importance
- Confidence Score: Returns the model's confidence alongside the predicted class
- REST API: FastAPI endpoint at /predict/burnout with JSON request/response
- Live Demo: Interactive slider UI on this page — try it without any signup
🧠
Try It Live
Enter your work profile — model predicts your burnout risk in real time
30yrs
2044
5yrs
019
8hrs
414
7hrs
39
2cups
07
5
019
8
029
3
09
10hrs
418
0.5hrs
02
50/100
0100
🎯
Adjust the sliders and click Assess Burnout Risk to see your prediction.
Technology Stack
Framework
scikit-learn + FastAPI
Platform
Web / Cloud
Language
Python 3.11
App Type
Health & Wellness AI
Tools & Services
PythonRandom ForestKNNImputerjoblibpandasNumPy
Ready to Build Something Like This?
Let's collaborate and bring your vision to life with production-grade engineering.