ai4i

End-to-end predictive maintenance project on the AI4I 2020 dataset. Includes feature engineering, calibrated logistic regression and random forest pipelines, with attention to class imbalance and operating-point evaluation (PR-AUC, recall @ 90% precision).

End-to-end predictive maintenance ML system with feature engineering, model comparison, FastAPI inference, automated testing, and CI.

Deep learning predictive maintenance project using a hybrid 1D CNN-LSTM model to classify machine failures from AI4I 2020 sensor data with TensorFlow/Keras and scikit-learn.

CPH4: an end-to-end predictive maintenance platform with Jupyter ML workflows, XGBoost failure-risk predictions, SHAP explanations, FastAPI, Next.js, batch CSV analysis, simulated telemetry, and Docker support.

MaintiQ Predict: Windows desktop MVP for smart manufacturing predictive maintenance with CSV risk prediction, SPC monitoring, GenAI manager reports, and approval workflow.

IP-safe competition evidence for IQFamily, IQCircuit's governed AI family-history platform

AssetIQ is an end-to-end AI prototype for predictive maintenance and asset intelligence. It helps maintenance teams in mining, energy, transport, telecoms, water and other asset-heavy industries decide which equipment needs attention first. I built this as an End-to-End AI Solution for various industries.

AI4I machine-failure risk triage from operating snapshots with engineered features, calibrated probabilities, Top-K ranking, FastAPI, and Streamlit.

Original-lineage replication materials for evaluator reliability in explanation-gated predictive maintenance; subsequent correction and verification materials are not yet included.