Data Scientist · MLOps Engineer
Models that reach production, and tell the truth.
I build the whole path — from the first notebook to a deployed, monitored API. Evaluated out-of-fold, so the number you see is the number you get.
Status
Open to roles
Role
Data Scientist · MLOps
Based
İstanbul
Selected work
01
LIVE DEMO6.49%
out-of-fold MAPE
Deployment · MLOps
Car Price Prediction & MLOps
End-to-end ML system on LightGBM with TF-IDF+SVD text features — scraping, dedup, leak-free 5-fold evaluation, drift monitoring, SHAP explainability, and a FastAPI serving layer. The complete production cycle, not just a notebook.
Technical arsenal
Grouped by where it earns its place — not a wall of logos.
Modelling & Stats
LightGBM · CatBoost
One leak-free 5-fold split — LightGBM won at 6.49% MAPE.
scikit-learn · TF-IDF+SVD
Listing titles into 170 dense dimensions, bundled with the model.
SciPy · statsmodels
Hedonic regression with bootstrap CIs; KS + Wasserstein for drift.
Serving & Data
FastAPI · DuckDB
Prediction, drift and BI endpoints aggregating 30K rows server-side.
Railway · S3
Model bundle and listing database load from object storage at boot.
pandas · NumPy
Scraping, dedup and the columnar aggregation behind the dashboard.
Frontend & Viz
Next.js
Static analysis report and the app pages rendering this work.
Plotly · ECharts
Notebook-grade figures and the live market dashboard.
Tailwind
The editorial and app layers of this design system.