I will build a python machine learning model for prediction and forecasting
Data Scientist, AI Engineer
About this Gig
Need a model that predicts something from your spreadsheet or database export? I build machine learning models in Python for classification, regression and forecasting, and explain what drives the predictions.
What I can do:
- Clean and prepare tabular data (CSV, Excel, SQL export)
- Train LightGBM, XGBoost, CatBoost or scikit-learn models with proper validation (no data leakage)
- Forecast sales, demand or user activity from historical data
- Explain predictions with SHAP: which factors matter, and how
- Review an existing model or notebook and find what's going wrong
Why me:
- ~4 years as a Data Scientist / AI Engineer
- Kaggle Silver Medal in a forecasting competition (rank 43/852)
- Built models and SHAP/PDP analysis on 20M+ records for a global pharma client
How it works: message me with a data sample and your goal. I confirm the scope before you order.
Results depend on data quality. I don't promise a specific metric, but I report exactly what was achieved and why.
Not included: dashboards, model deployment or APIs, data collection or scraping.
Programming language:
Python
•
SQL
•
Colab
Frameworks:
Scikit-learn
•
Google ML Kit
•
Panda
Tools:
Jupyter Notebook
•
Excel
•
MLflow
•
Colab
My Portfolio
Other Data Science & ML Services I Offer
FAQ
What data formats and sizes do you accept?
CSV, Excel, Parquet or a SQL export. One table, or a few tables that join on a clear key. Up to 100,000 rows and 50 columns; larger data needs a custom offer.
What will I receive?
Basic: a written report. Standard: Jupyter notebook, trained model file and metrics on held-out data. Premium: Standard plus a comparison of two approaches, SHAP plots and a written summary of the key drivers.
Can you forecast many products or stores at once?
Standard and Premium cover one dataset with one target. Forecasting many separate series at once (for example hundreds of products or stores) needs a custom offer, so please message me first.
Do you deploy the model?
No. You get a notebook that loads the model and makes predictions on new data. Deployment as an API or on a cloud service is not included.

