Looks Like This Service Is On Hold
I will build a neural network in pure python without libaries
About this Gig
I build neural networks using only Python + NumPy. No TensorFlow. No PyTorch.
Most freelancers call model.fit() and hand you a black box. I write every gradient and weight update from scratch. You get code you can read, audit, and own.
Perfect for: fraud detection, churn prediction, price forecasting, risk scoring.
Requirements: tabular data (CSV), clear target column or another adequate data format
I do NOT work with: images, text, or TensorFlow.
Message me before ordering with: what you're predicting, rows/columns, target accuracy.
Expertise:
Feature learning
•
Classification
•
Predictive analysis
Programming language:
Python
Frameworks:
Other
APIs:
Other
Tools:
Other
FAQ
Do you use TensorFlow or PyTorch?
No. I build everything from scratch using only Python and NumPy.
What problems can your MLP solve?
Binary classification, multi-class classification, and regression.
How much data do I need?
Basic: 100+ rows. Standard: 1,000+ rows. Premium: 5,000+ rows.
What do I get at the end?
Python script with your MLP, trained weights, prediction function, and documentation.
Can I use this commercially?
Of course!
Do you work with images or text?
No. Tabular data only.

