Data Scientist · MLOps Engineer

Building intelligent data solutions, end to end.

Bridging the gap between data science and production — the models, the pipelines, and the dashboards that make them useful.

Used-car prices, 2005–2025SAMPLE FIGURE

Selected metrics across recent work

13,904
Rows modelled
0.98
Best holdout R²
7.4%
Overall MAPE
₺123K
Overall MAE
live API
Deployment · MLOpsLIVE DEMO

Car Price Prediction & MLOps

End-to-end ML system on CatBoost with MultiQuantile loss — data ingestion, training, drift monitoring, SHAP explainability, and a serving API. The complete production cycle, not just a notebook.

CatBoost · FastAPI · Hugging Face

Technical arsenal

Not a list of logos — what each tool actually did, across modelling, analysis, and production.

CatBoost · scikit-learnGradient boosting for price prediction; classical models where interpretability mattered more than raw accuracy.
pandas · scipy95% confidence-interval methodology over 70k+ rows — statistical rigor before reaching for ML.
FastAPIVersioned inference endpoint serving predictions in real time.
Hugging Face · RailwayModel registry and zero-config deployment behind a promote-to-production flow.
ECharts · Next.jsThe live dashboards and figures rendering this work in the browser.

Latest writing

All posts →
2026-01-28End-to-End Car Price Prediction System and MLOps Architecture