A semantic search engine powered by sentence transformers that understands natural language queries and retrieves relevant documents with high precision.

• Sub-100ms query latency at scale
• Supports 12 languages
• Open-sourced with 2k+ GitHub stars

An interactive web application for experimenting with language models, comparing outputs, and visualizing attention patterns in real time.

• Side-by-side model comparison
• Real-time token streaming
• Attention visualization

A lightweight annotation tool designed for small research teams, supporting text classification, NER, and relation extraction with built-in quality control.

• Multi-annotator agreement tracking
• Active learning suggestions
• Export to common ML formats

A Telegram bot that summarizes academic papers from arXiv, extracts key contributions, and generates concise reading notes automatically.

• Processes 50+ papers daily
• Structured summary format
• 500+ active users