Showing 30 of 1,490 · Category: Machine Learning
ML-powered FastAPI service predicting student mental-health score from social-media usage, sleep, study & stress. Random Forest pipeline (R² 0.88), Pydantic validation, Swagger UI.
Fraud-network investigation prototype: analysts triage risk alerts, trace money flows between accounts on an interactive graph and score transactions with a Random Forest model. Next.js · Supabase · FastAPI · scikit-learn. Synthetic PaySim data.
Turn raw workflow records into strict JSON. Give it an issue, a log line, a commit message, a pull request, an AI agent tool call, a compiler diagnostic, a git diff, or a code review comment - or plain non-workflow text. Get a validated JSON object.
Telecom contact center fraud detection and customer churn prediction system (2010–2014)
CS229 notes, CS229 lecture notes — CS229 2018 machine learning lecture notes
Production-grade AI Agent reference for any coding agent (Claude Code / Cursor / Copilot / Codex). 24 reference nodes: ReAct Loop, MCP, RAG, Multi-Agent, Memory, Guardrails. Includes build-vs-buy decision tree + integration patterns. No deprecated stack, all TypeScript production code. https://morningstar202604.github.io/agent-builder
Neural localization of photon ring structure in synthetically degraded black hole images via U-Net
Top Distributed AI Computing (Opensource) Star if you like it!
Fine-Grained Evaluation of Human Gameplay Behaviours in Rocket League
Admission control and queueing simulator for LLM inference: backpressure, priority scheduling, capacity planning and tail latency
Hierarchical budget enforcement and burn-rate alerting for LLM spend, with a concurrent reserve/commit/refund ledger
Deadline-aware LLM gateway: shared request deadlines, bounded retries, circuit breakers, fallback and cost-amplification analysis
Everything useful for predicting magnetic materials and their properties. An open collection of tools, datasets, machine learning models, simulation frameworks, and visualization software for magnetic materials research.
Delivery intelligence platform with late-risk scoring, FastAPI, PostgreSQL, and Streamlit
🚀 The ultimate AI agent framework using the Model Context Protocol (MCP). Bridge Python and Node.js to instantly automate tasks, build custom bots, and connect your favorite AI models to real-world developer tools.
ML Project, trying to implement the theoretical knowledge I obtained through Google's ML Course.
CP asymmetry in B± → h±h+h− decays with LHCb Open Data: frequentist, Bayesian and ML analysis
Monitoreo agroclimático en tiempo real: MQTT, Node-RED, InfluxDB y machine learning en línea.
Sparse connectome models for custom task training and stateful inference. MaleCNS graph bindings, local PyTorch adaptation, and portable checkpoints.
End-to-end retail analytics platform — Python, SQL, RFM segmentation, forecasting, Streamlit dashboard, Power BI & Tableau
AI-powered real estate intelligence platform for property valuation, smart recommendations, market insights, document analysis, and personalized AI assistance.
Mall Customer Segmentation using Machine Learning | An unsupervised machine learning project that analyzes customer data to identify groups with similar characteristics and spending behavior. The project explores customer segmentation, data visualization, clustering concepts, and insights that can support customer
Privacy-preserving federated learning in NumPy: FedAvg (IID/non-IID), Poisson DP-SGD with a subsampled-Gaussian RDP accountant, secure aggregation, and membership-inference attacks measuring real leakage vs epsilon.
CNN-based handwritten digit recognition using TensorFlow/Keras and the MNIST dataset.
End-to-end ML systems: Titanic survival and heart disease prediction with SVC (80% and 75% test accuracy), full Jupyter notebook and a deployed two-tab Streamlit app.
A clean LaTeX template for machine learning papers, with a customizable title page, resource links, and an organized appendix.
Knowledge tracing on ASSISTments: baselines, IRT, PFA and DKT with leak-free evaluation.
Open benchmark for credit-default prediction with scarce and delayed labels: 19 public datasets, label-delayed streaming and i.i.d. evaluation, leakage-controlled synthetic augmentation. pip install pamir
Modern Sudoku game solved in real-time by Laya System 1 decision model (convaiinnovations/laya) with ~35ms GPU latency.