LLM Router is a service that can be deployed on‑premises or in the cloud. It adds a layer between any application and the LLM provider. In real time it controls traffic, distributes a load among providers of a specific LLM, and enables analysis of outgoing requests from a security perspective (masking, anonymization, prohibited content).

A hybrid Relational-GAT with a projection head that learns relation operators to transform embeddings, impute unseen node vectors, and expand search queries via relational paths.

A lightweight, extensible toolkit for preparing, indexing, and querying textual data for semantic search and downstream NLP tasks.

A companion repository for llm-router containing a collection of pipeline-ready plugins. Features a masking interface for anonymizing sensitive data and a guardrail system for validating input/output safety against defined policy rules.

Utilities and scripts to transform plWordNet into machine‑learning‑ready artifacts: build graphs, create balanced embedding datasets from lexical relations, split into train/test, and prepare indices for downstream retrieval tasks.

A lightweight, configurable toolkit for unsupervised clustering of short or long texts. It builds vector embeddings for your data, discovers dense regions with HDBSCAN, optionally reduces dimensionality for visualization (e.g., t‑SNE), and exposes graph-friendly outputs for downstream exploration.

This repository contains a collection of LaTeX code, primarily focused on demonstrating and exploring various concepts and techniques. You'll find detailed descriptions of our playground (playground.radlab.dev) and explanations of different methods and approaches developed by us.

A collection of helper utilities and handlers designed to simplify common machine‑learning workflows.

A Polish extractive Question‑Answering (QA) solution with utilities for mapping its token space to generative LLMs (Gemma, LLaMA, GPT‑OSS). The repo contains scripts for training the QA model, inference pipelines, a CLI to pre‑compute token‑mapping tables, and a generation helper that injects static bias from QA logits into a downstream generator.