Comments & Twitter accounts gRPC classification service.

A reproducible evaluation pipeline for assessing toxic content in language model outputs using Detoxify, developed as part of a 10-week research project.

Reproduction study of the paper "Detoxifying Text with MaRCo: Controllable Revision with Experts and Anti-Experts"

A simple FastAPI wrapper around the Detoxify library

Bachelor's thesis on removing hate from online comments using paraphrasing: algorithm DPhate

AI-based video moderation system that detects violence and toxic speech using multimodal analysis and applies selective content filtering with policy-based reporting.

Model-agnostic fast api service to evaluate large language models for bias, toxicity and hallucinations.