Weight Your Wage

Predicts developer salaries from the Stack Overflow survey. A 42AI AI Lab team project covering the full path from raw data to a public web app; the model is a deliberate baseline, the pipeline is the work.

TL;DR

Deployed · public web app
Team of 5

My partMy commits went mostly to the model training code (PyTorch Lightning, MLflow), the FastAPI service, the Docker Compose setup and the monitoring dashboards.

Code ↗Demo ↗

ROLE: Team of 5
CONTEXT: 42AI AI Lab
STATUS: Done
STACK: Python PyTorch PyTorch Lightning pandas scikit-learn Cleanlab MLflow FastAPI PostgreSQL MinIO Docker Docker Compose Prometheus Grafana Loki
RESULT: Deployed and public. No accuracy claim: the model is a placeholder by design.
LAST UPDATED: 26 Sep 2026

Data

Stack Overflow Developer Survey 2025.

What I built

Team of 5

My commits went mostly to the model training code (PyTorch Lightning, MLflow), the FastAPI service, the Docker Compose setup and the monitoring dashboards.
I built the Next.js web front end alone, in a separate repository.

raw survey

MinIO

ETL

PostgreSQL

clean labels

Cleanlab

train · Huber

Lightning · MLflow

serve

FastAPI jobs

web app

Next.js

monitoring — Grafana · Prometheus · Loki

Fig. 1 — Pipeline as implemented in the repo. Stages outlined in accent are where most of my commits went.

Key choices

Huber loss: Less sensitive than squared error to the extreme salaries found in survey answers.
Cleanlab before training: Flags survey rows whose labels look wrong before the model sees them.
Predictions as async jobs: The API returns a job and the web app follows its progress instead of blocking on the model.

Results

Deployed and public. No accuracy claim: the model is a placeholder by design.

Limits

The model is a placeholder: the project measures the pipeline, not prediction quality.
One survey year.
The prediction API is currently offline; the web app still loads.

Questions about this project? → Email me