Vahdettin Karatas
Freelance ML / data systems — pilots and internal tooling
  • Location:
    Prague, Czech Republic
Technical focus
  • Churn and retention analytics pilots
  • JSON APIs for operational workflows
  • Feature preprocessing contracts
  • Deployment with clear data-handling boundaries
  • Honest scoping—evaluation vs production hardening
Capability overview — prioritization

Customer churn risk scoring

Score individual accounts from structured fields (Telco-style schema): JSON in, churn probability plus coarse risk tier out—suited to trials where teams want ranked follow-up, not a replacement for CRM strategy.

Use Open scoring demo (API) on churn-api for the full interactive UI. Governance caveats still apply.

Python
scikit-learn
FastAPI
Docker

Limitations & scope

  • Demonstration training uses the public Telco Customer Churn dataset; useful scores on your accounts require retraining and validation on your own labeled data.
  • The HTTP API exposes POST /predict with one JSON record per request. There is no server-side CSV upload or batch queue in this codebase.
  • Risk bands (high / medium / low) use fixed probability cutoffs for UI grouping; the binary churn label uses a threshold read from training outputs (see metrics).
  • The published Docker image is inference-only: you mount a trained joblib artifact plus metric files from the training pipeline (see README).
  • Risk bands do not encode business cost matrices or uplift models.

Try the scorer

The interactive demo (metrics strip, findings, model chart slot, customer form) lives on churn-api — same presentation as the shipped Vercel-style demo; this labs page stays positioning-only.

Production fit and integration remain separate from what you see in the browser.

Churn risk scoring

JSON API · probability & tiers · pilot framing

© Vahdettin Karatas. All rights reserved.