LogIQ v0.1.0 – Initial Hybrid Log Classification System

Initial release of LogIQ – a hybrid log classification system combining rule-based filtering, semantic embeddings, and machine learning for intelligent log categorization.

Key Features

• Hybrid classification pipeline using Regex rules, Sentence Transformers embeddings, and a Logistic Regression classifier
• Confidence-based routing with LLM fallback for ambiguous logs
• FastAPI inference service for real-time log classification
• Modular architecture separating preprocessing, embeddings, and classification layers
• Reduced unnecessary LLM usage by ~70–80% through confidence-based routing

Tech Stack

Python, FastAPI, Sentence Transformers, Scikit-learn

This release introduces the first working version of the LogIQ system capable of classifying logs through a hybrid ML + LLM pipeline.