Door categorieën bladeren
Ontdekken
Fiverr Pro
Nederlands
$
USD
I will build machine learning and predictive models
Over deze dienst
Need a machine learning model that actually works on your data, not just in a notebook?
I'm a Data Scientist with an MSc in Data Science (UK) who builds, evaluates and deploys ML models in Python. I focus on results you can trust: clean data, leakage-free testing and the right metrics for your problem.
WHAT I CAN BUILD FOR YOU
- Customer churn prediction
- Sales and demand forecasting
- Classification and regression models
- Clustering and customer segmentation
- Anomaly detection
- Live web apps or APIs for your model (Premium)
WHAT YOU GET
- Data cleaning and exploratory analysis
- Model built with Scikit-learn, XGBoost or TensorFlow/Keras
- Clear results report in plain English
- Full Python source code and Jupyter notebook
RECENT WORK
- Telecom churn model: 4 models compared, AUC 0.829, deployed on AWS
- Predictive maintenance: GRU model forecasting engine failure, deployed on Google Cloud
- Income classifier: 87% accuracy, 0.92 ROC-AUC
Please message me BEFORE ordering with a short description of your data and goal, so I can confirm the right package for you.
Programmeertaal:
Python
•
R
•
SQL
•
Colab
Frameworks:
Scikit-learn
•
keras
•
PyTorch
•
Panda
•
Overige
Tools:
Jupyter-notitieboek
•
tensorflow
•
Colab
•
RStudio
Veelgestelde vragen
What do you need from me to start?
Your dataset (CSV, Excel or a database export), a short description of what you want to predict, and any deadline or tool preferences. If your data isn't ready yet, message me and I'll help you plan what to collect.
Which package should I choose?
Basic if you need one model tested on one dataset. Standard if you want the best of several models, tuned and explained. Premium if you need the model running as a live app or API that your team can use.
Can you guarantee a specific accuracy?
No honest data scientist can guarantee accuracy before seeing the data, because it depends on data quality and how predictable the problem is. I'll tell you early what results look realistic, and I always report results on unseen test data, not inflated training scores.
Will I get the source code?
Yes. Every package includes the full Python source code and a Jupyter notebook, so you own everything and can rerun or extend the model.
Is my data kept confidential?
Yes. I use your data only for your project, never share it, and delete it after delivery if you ask. I'm happy to sign an NDA.
