Skip to content
View josortmel's full-sized avatar
2 followers · 0 following

Block or report josortmel

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
josortmel/README.md

José Antonio Ortiz Melo — AI engineer · multi-agent systems

EcoDB — collective AI memory infrastructure, in motion

I build memory infrastructure for AI agents, and I direct teams of agents to ship it. Based in Sevilla.

EcoDB and the projects below were built the way I work: one human holding the architecture and the calls, directing specialized AI agents (design, backend, QA, security) that each own their craft. I care about systems that hold up in production, not demos: real benchmarks, real isolation, real governance.

What I've built

EcoDB — collective AI memory infrastructure

Shared memory for teams of agents. A 10-stage retrieval pipeline (GAMR) over a knowledge graph in PostgreSQL, with cross-modal search (a text query finds images and vice versa), multi-tenant isolation, role-based governance, and a desktop dashboard.

  • Recall@5 of 0.92 on LoCoMo (ACL 2024); full GAMR pipeline under 50 ms p95
  • Knowledge graph on Apache AGE (Cypher inside Postgres), ~100 canonical predicates
  • 32 MCP tools + 30+ REST endpoints; in production since May 2026

Python · FastAPI · PostgreSQL/pgvector · Apache AGE · Docker · Jina embeddings · MCP · React · Electron

Eco Relay — real-time messaging between agent sessions

Direct messages, rooms, and presence so independent AI agent sessions can coordinate on the same machine. Persistent delivery, IRC-style ephemeral rooms.

Stack

Backend: Python, FastAPI, PostgreSQL (pgvector, Apache AGE), Docker. AI: embeddings (Jina v4), GLiNER NER, cross-encoder reranking, MCP. Frontend: TypeScript, React, Electron, Tailwind.

Reach me

Popular repositories Loading

  1. EcoDB EcoDB Public

    Collective AI memory for multi-agent teams. 10-stage GAMR search engine, knowledge graph (Apache AGE), document ingestion, 32+ MCP tools.

    HTML 8

  2. EcoRelay EcoRelay Public

    Inter-session messaging for AI coding assistants. Hub-and-spoke architecture with persistent groups and LAN federation.

    TypeScript 3

  3. obsidian-vault-indexer-for-ai obsidian-vault-indexer-for-ai Public

    # AI Vault Indexer for Obsidian This plugin automatically generates a comprehensive index of your Obsidian vault, specifically formatted for **AI agents** and **LLMs** (Large Language Models).

    1

  4. Faro Faro Public

    Multi-agent orchestration system for Claude Code CLI. 24 agents, 12 workflows, 90+ days of production usage.

    Python 1

  5. obsidian-mcp-tools obsidian-mcp-tools Public archive

    Forked from jacksteamdev/obsidian-mcp-tools

    Pretty similar to the original repo with the only difference that it has a new tool read_note_range that allow for parcial reading of a document.

    TypeScript

  6. josortmel josortmel Public