Welcome to my Universe
Where raw data transforms into actionable intelligence. Explore machine learning models, statistical architectures, and client-side predictive computing deployed with zero latency.
Real-time Predictive Inference Engine
Simulate an in-browser predictive regression model. Adjust the inputs below to calculate real-time estimated valuation with instant statistical confidence intervals.
Specialized Machine Learning Spokes
Each specialized model runs in isolation on its own high-speed edge page without taxing your CPU or backend servers.
House Price Predictor
Evaluates multi-variable property metrics (square footage, bedrooms, zip code location) with instant client-side valuation.
Stock Trend Forecaster
Computes moving averages, volatility bands, and predictive directional momentum without sending ticker data to any server.
Credit Loan Classifier
Predicts applicant risk profile and loan approval thresholds using optimized decision boundary trees.
Customer Churn Neural Net
Multi-layer perceptron calculating customer retention probabilities with client-side privacy preservation.
The Zero-Server Pipeline: From Jupyter Lab to Cloudflare Edge
Train & Export
Models are trained in Jupyter Lab with Scikit-Learn or PyTorch, then serialized into open ONNX computational graph binaries.
WebAssembly Runtime
The browser loads `onnxruntime-web` to execute the mathematical graph directly using CPU threads with sub-2ms latency.
Cloudflare Edge
All static assets, scripts, and models are served from 300+ global Cloudflare edge locations with instantaneous worldwide delivery.
Frequently Asked Questions
What makes this Data Science Universe unique?
Every model in this universe is engineered to run directly inside the visitor's browser using ONNX Runtime and WebAssembly. There are zero backend server spin-up delays, zero cold starts, and your data never leaves your device.
How does the machine learning model run without a Python backend?
After training in Jupyter Lab using Scikit-Learn or PyTorch, models are exported into the open ONNX standard. Modern browsers execute these mathematical graphs using WebAssembly (WASM) at near-native C++ performance.
Can new models be added to this universe easily?
Yes! Using our modular Hub-and-Spoke structure, each trained model is isolated into its own dedicated spoke route, sharing the same lightweight global design system.