Exploring possible methods for Audio Anomaly Detection - on machine sounds (MIMII dataset)

Project on implementation of XAI in wearables using a dataset provided by Kaggle

[Anomaly detection] refers to the process of identifying patterns in data that do not conform to expected behavior. This project aims to develop a machine learning model to predict and identify potential attacks in IoT networks, thus helping to secure these networks from malicious activities.

Group 13's project for the WASP course: Scalable Data Science and Distributed Machine Learning

Flow-level behavioural detection of command-and-control beaconing under timing jitter, size variation, burst traffic, hard benign profiles, and CTU-13 public-data domain shift. Includes synthetic benchmarking, interpretable/statistical/anomaly/supervised baselines, minimum-evidence analysis, CTU-native validation, and report-ready results.

🚀 Detect anomalies in structured datasets with this AI-driven ETL pipeline, ensuring data quality through seamless ingestion and machine learning insights.

Agentic Data Engineering Platform is an open-source, production-ready ETL solution that combines the Medallion Architecture with AI-powered agents that autonomously profile, clean, and optimize your data—so you can focus on insights, not infrastructure.

AI-powered ETL pipeline with ML anomaly detection and FastAPI deployment

Benchmarking Temporal Convolutional Networks (TCN) vs RNN, LSTM, and GRU for long-range sequence modeling with PyTorch; includes modular framework, stability improvements, and empirical analysis.

A predictive maintenance framework contrasting feed-forward and sequential machine learning models to forecast industrial machinery failures, achieving near-perfect AUROC scores with GRU models.

Data pipeline and analytics toolkit for Garmin smartwatch data using GarminDB exports. Extracts, aggregates, and cleans health metrics in SQLite, with tools for modeling, anomaly detection, and visualization.

BigFoot is a comprehensive analytics tool designed for the analysis of Bigfoot sighting data, using machine learning and data visualisation. The application imports verified sighting reports from the BFRO (Bigfoot Field Researchers Organisation) and provides interactive dashboards, maps, and predictive tools for exploring cryptid reports.

Autonomous home network monitoring platform — 10-phase AI audit pipeline, CLIde AI-verifying-AI architecture, 5-stage syslog compression (99.97% reduction), predictive analytics, 17 Discord chat tools

Machine learning based intrusion detection system

DeepTrace - Real-time observability layer for agentic AI systems. Intercept, trace, visualize, and secure every LLM inference and tool invocation across your agent swarm.