Real-world time series data from domains such as environmental monitoring and industrial systems are often irregularly sampled and multivariate, challenging the assumptions of traditional time series analysis methods. Most existing approaches rely on regular sampling and univariate inputs, limiting their ability to capture interactions across variables and time. These constraints impede the discovery of motifs, recurrent patterns revealing system behavior, and discords, which signal anomalies or faults. We propose AMPIIMTS (Adaptive Matrix Profile Indexing for Irregular Multivariate Time Series), a matrix-profile-based framework for efficient motif discovery and anomaly detection in such settings. AMPIIMTS extends matrix profile computation to irregularly sampled, multivariate data, enabling robust pattern extraction under real-world conditions. The proposed end-to-end pipeline ingests raw time series, performs adaptive normalization and alignment, computes cluster-aware matrix profiles, and automatically identifies meaningful motifs and discords. This work bridges advanced time series methodology with practical usability through a lightweight Python toolkit that abstracts algorithmic complexity and supports scalable analysis of challenging datasets.