Continuous, fine-grained water-quality monitoring is essential to prevent disease outbreaks and optimize yields in intensive shrimp aquaculture. We present an end-to-end IoT \(\rightarrow \) ETL \(\rightarrow \) AI framework that integrates 14 in-situ sensors, hourly cleansing and feature extraction, and sliding \(10\times 14\) windows. Building on prior work, we adopt autoencoder-based anomaly detection trained only on normal data and contrast it with sequence forecasting, reflecting the trade-off often noted in aquaculture time-series literature between sensitivity and complexity. Our method instantiates three candidates: a 1D CNN Autoencoder, a Fully-Connected Sequence Autoencoder, and an LSTM Forecaster; anomaly thresholds are set from validation errors using \(\mu +2\sigma \) . On a one-month dataset from an intensive shrimp farm, we apply stratified splits for train/validation/test. Results show the CNN–AE delivers the best precision–recall balance for early warning; FC–Seq–AE preserves high precision with reduced recall and smaller footprint; LSTM–Forecaster achieves zero false alarms but misses many events. The pipeline includes defense-in-depth security (mTLS-protected MQTT, AES-256 at rest, RBAC) and can support federated learning. Discussion highlights limitations (single site, early/mid-season coverage) and the need for multi-farm, late-season data and cost-sensitive learning to further improve recall and generalization.