Distributed AI platform for crypto analysis. A Node.js scraper ingests news into a Redis queue, feeding a Go LLM service that uses OpenAI function-calling for Binance data and Qdrant vector embeddings for semantic retrieval. Output: structured trade positions (entry, stop-loss, take-profit, rationale) with live P&L. Next.js dashboard + companion Terminal UI. Deployed on Hetzner + Vercel.

• Node.js scraper pulls articles from multiple crypto news sources on a cron schedule.
• Go LLM service calls OpenAI with function-calling for live Binance market data.
• RAG pipeline using LangChain embeddings stored in Qdrant for semantic context.
• Structured trade positions persisted in Postgres and streamed via WebSocket.
• End-to-end observability with Prometheus, Grafana, OpenTelemetry, and Jaeger.

A real-time virtual classroom platform enabling live teaching sessions with synchronized video, whiteboard, and PDF sharing.

• Live video streaming.
• Synchronized whiteboard drawing, stroke customization, and instant updates across all participants.
• PDF presentation with synchronized navigation.
• Permission-based access and participant removal capabilities.

A terminal-based interface for Alpstein providing real-time crypto insights and trading analytics directly in your command line. Built for traders who prefer keyboard-driven workflows and minimal context switching.

• Real-time market updates with live price feeds and trading signals streamed to your terminal.
• Agent opinion display with position recommendations, risk/reward ratios, and live P&L tracking.
• Authentication flow with secure auth key input for accessing the Alpstein platform.
• Clean TUI design using Bubble Tea's component model for responsive, interactive terminal experience.