Browse categories
Explore
Fiverr Pro
English
$
USD
I will train custom ml models, fine tune llms, gpt and llama on your data
AI Engineer, LLM Agents, RAG and n8n Automation
I build AI agents, RAG systems and n8n automations that replace manual work in real businesses.
• AI agents & RAG: LangChain, LangGraph, LlamaIndex, pgvector
• n8n workflows: CRM, email, e-commerce, ...
About this Gig
Have data but no clear way to turn it into predictions? I train machine learning models on your data and deliver them ready to use, not just a notebook.
WHAT I BUILD
- Prediction and forecasting: sales, demand, prices, time series
- Classification: churn, lead scoring, fraud, risk, text categories
- Regression for business metrics
- Clustering and customer segmentation
- Anomaly detection and recommendation systems
- Text models: sentiment and topic classification
HOW I WORK
1. Data review: quality check, missing values, leakage risks
2. Feature engineering and a solid baseline
3. Compare algorithms (XGBoost, LightGBM, scikit-learn, PyTorch) and tune
4. Honest evaluation on held-out data with clear metrics
5. Explain the drivers behind predictions (SHAP)
6. Deliver the model with a FastAPI endpoint and Docker, if needed
YOU GET
- Trained model file and clean, commented code
- Metrics report in plain language
- Prediction script or API for real use
- Documentation and handover
Not sure ML fits your problem? Send me a short description and a data sample, and I will tell you what accuracy is realistic before you order.
Programming language:
Python
•
SQL
Frameworks:
Scikit-learn
•
Google ML Kit
•
Keras
•
PyTorch
•
Panda
APIs:
Microsoft Computer Vision AI
•
Google Cloud Vision API
Tools:
Jupyter Notebook
•
OpenCV
•
TensorFlow
•
MLflow
•
Colab
My Portfolio
FAQ
What data do you need?
A CSV, Excel file or database export with historical records and the column you want to predict. A small sample is enough to assess feasibility before ordering. More quality data usually means a better model.
What accuracy can I expect?
It depends on your data. After a quick data review I give a realistic estimate, then report honest metrics on held-out data (accuracy, F1, ROC-AUC, MAE or RMSE). I never inflate results with data leakage.
Do I get a model I can actually use?
Yes. You get the trained model file, a prediction script and, in Premium, a FastAPI endpoint in Docker that your app, website or n8n workflow can call.
Can you work with images or text?
Yes. Text classification and sentiment are included. Image classification and object detection (YOLO) are available as a custom offer after we review your dataset.
Is my data kept confidential?
Yes. Your data is used only for your project, is never shared and is deleted after delivery on request. I am open to signing an NDA.
Can you retrain the model later on new data?
Yes. I can add a retraining script as an extra, so you can update the model yourself, or retrain it for you in a follow-up order.
