AI-powered predictive maintenance system using Python, Machine Learning, and simulated IoT sensor data to predict equipment failure and visualize risk through an interactive dashboard.

Biscuit defect classification using Classical Machine Learning, TinyVGG CNN, and EfficientNet-B0 Transfer Learning.

Machine learning project for customer churn prediction using Logistic Regression, Decision Tree, Random Forest and Gradient Boosting with model evaluation and visualization.

A self-learning Python predictor that analyzes human number-selection patterns using adaptive statistical models, persistent profiles, and real-time accuracy tracking.

Ink into Stars: 108 Lectures on Machine Learning | 《墨点成星：机器学习一百零八讲》中英双语讲义与配套代码

Employee attrition segmentation on IBM HR data with K-Means, GMM, Agglomerative & K-Prototypes + PCA | Çalışan devir hızı kümeleme analizi

Health Insurance Claim Fraud Risk Scoring System: Pendekatan statistik dan pembelajaran mesin (LightGBM & CatBoost) untuk mengesan, mengkuantiti, dan mengutamakan risiko penipuan tuntutan insurans kesihatan di bawah kapasiti audit yang terhad (5%) menggunakan metrik utama NormalizedRecall@5%.

Lightweight open-source AI training platform on Kubernetes: submit Python scripts as batch Jobs, queue with priorities and quotas, track curves, register models, run batch and online inference. No CRDs. | 轻量开源 K8s AI 训练平台

Free, open-source AI photo editing for Lightroom Classic. Learns your editing style, applies presets/LUTs and AI masks locally. An Imagen AI and Aftershoot alternative.

Structure-based drug discovery toolkit that makes AI models easy for scientists to use: pocket detection, structure prep, and binder generation.

Open-source, local-first YOLO annotation tool. Label bounding boxes and polygons in a browser on your own computer, import and export YOLO datasets, and work in a dockable workspace. English by default, 12 languages, no account. AGPL-3.0.

A map of the places gradient descent gets stuck, and the ways it gets out. Interactive loss surfaces from the optimization literature, each citing its sources.

View Engine — local AI research and trading desk: an orchestrator hires agent teams on your Claude Code subscription, plus RLCD, a calibrated buy/sell/hold model, and annotated charts for gold and FX.

Alefba Reader: reads Persian letters from a photo (printed, handwritten, coloured, tilted or sideways, marked on a sheet, or several in a row) and gives their romanised spelling. OpenCV + two PyTorch CNNs, Flask web app for laptop and phone with an alphabet chart, dark mode and the Noghte mascot. CIT 401A-IT31S1, TIP Manila.

Which prototype aggregation rules resist collapse in continual crop-disease learning: HDC vs NCM, RanPAC-style, Sign/Median-NCM on DINOv3 features

Tech internships open to students in India, updated daily. SWE, ML, data, quant. Run by IITians, open to every student.

VS Code extension to review Jupyter notebooks (.ipynb) in GitHub pull requests: rendered cell-by-cell diffs of code, Markdown, LaTeX and outputs, with review comments on any cell that post to GitHub.

Systematic quant trading platform on the IG API: streaming data, quant methods, C++ HFT engine, ML signals, LangGraph agents.

Fast, local language identification for 100 languages with a compact PyTorch model.

Multi-Gameweek Fantasy Premier League planner: search over transfers, XI, captain and chips under uncertain player forecasts, backtested against greedy and rule-based baselines.

Tajik instruction-tuning dataset for LLMs | Датасети instruction-tuning забони тоҷикӣ (NLP, low-resource)

Monthly crime prediction per LSOA from 14.8M police.uk records and Census 2021: OLS vs Random Forest vs XGBoost

Inference code to run GPT-2_Instruct-v0.6

End-to-end ML pipeline predicting machine failures on the AI4I 2020 dataset (imbalanced classification).

Abalone yaş grubu sınıflandırması: 7 denetimli ML modeli (LR, KNN, karar ağacı, bagging, RF, XGBoost), CV tabanlı model seçimi ve SHAP ile açıklanabilirlik

Computational biology research prototype for microalgal culture vulnerability assessment using transcriptomics, statistical analysis, ML, biological networks, and economic decision support. No biologically validated pathogen-risk model yet.

Track 3: DFIR telemetry extraction, Shannon entropy profiling, and ML anomaly classification for CI/CD build logs.

Physics-informed neural networks for forward and inverse heat, Burgers, and Navier–Stokes problems using PyTorch and DeepXDE.

Differentially private CNN on MNIST with PyTorch + Opacus DP-SGD: 90.0% test accuracy at ε = 0.30 (δ = 1e-5).

An interactive visual journey through convolution, feature maps, pooling and prediction in CNNs. Runs entirely in the browser.