Building intelligent systems that reason, collaborate, and scale
We're a Barcelona-based team exploring the frontiers of multi-agent systems and AI engineering. We build production-grade systems and share what we learn along the way.
Multi-Agent Systems: Orchestrating autonomous agents with LangGraph, implementing Model Context Protocol (MCP) servers, and exploring emergent behaviors in agent communication
AI Infrastructure: Production-ready observability pipelines, token optimization strategies, and distributed inference systems
Research → Production: Turning AI research into working systems. From RAG with hybrid search to fine-tuning pipelines and custom embedding models
Agent Orchestration: LangGraph · CrewAI · Custom MCP Servers
LLMOps: LangSmith · ORQ · Prometheus
Vector & Search: Pinecone
Languages: TypeScript · Python
Frameworks: SvelteKit
We're always interested in connecting with fellow AI engineers, researchers, and builders.
- Website: eigensource.com (yes, it's a terminal interface)
- GitHub: You're already here! Check out our repos
- Twitter: @eigensource
- Email: info@eigensource.com
Building from Barcelona · Thinking in computational graphs