Founder of Sinensia — IT training and infrastructure services from Spain.
I build production systems, teach how they work, and ship open-source tools along the way. My focus sits at the intersection of security engineering, AI/agent infrastructure, and sovereign computing.
DevSecOps & SIEM Long background with IBM QRadar and IBM Guardium from earlier consulting years, complemented by Elastic / ELK, Cribl Stream pipelines, and the CI/CD tooling around them (Jenkins, Ansible, Terraform). Comfortable across the full chain: from raw log ingestion and parsing, through correlation rules and offense management, to dashboards and stakeholder reporting.
AI / ML & Agent Infrastructure Building and teaching the practical side of LLM-powered systems: MCP servers, Skills, agent memory, retrieval, evaluation, and observability with tools like Langfuse. Hands-on with PyTorch, scikit-learn, and the standard data stack (pandas, NumPy, Matplotlib) for cybersecurity-focused ML use cases — anomaly detection, fraud, attack classification.
Ethical Hacking & Offensive Security Designing and delivering ethical hacking courses with full lab infrastructure (Kali, DVWA, custom CTF-style targets). Walkthroughs cover SQLi, XSS, RCE, privilege escalation, AD attacks, and network forensics with Wireshark, Nmap, Metasploit, Burp Suite.
Cloud-Native & Kubernetes Running production workloads on k3s over bare-metal, with cert-manager, Traefik, Velero backups, NetworkPolicies, Kustomize, and a deterministic CI/CD flow. Migrated from Docker Swarm to k3s without downtime on live multi-tenant services.
Bitcoin & Sovereign Computing Long-term Bitcoin holder and full-node operator. Following post-quantum cryptography developments closely (Grover vs Shor, BIP-360/361, ECDSA exposure timelines) and exploring how decentralized payments fit into agent-to-agent economies.
Graph-based long-term memory framework for AI agents. Stores entities, observations, and relationships in a Neo4j knowledge graph, exposed to any MCP-compatible client. Apache 2.0, with a plug-and-play open-core architecture and an optional managed cloud sync. Stack: Python, FastMCP, Neo4j, Ollama (optional embeddings).
Open-source decentralized protocol for AI agents combining post-quantum cryptography, Bitcoin / Lightning payments, and censorship-resistant P2P networking (Kademlia DHT). Goal: two agents can discover each other, negotiate a task, and settle payment with no central infrastructure. Stack: Rust workspace, libp2p, LDK.
A growing set of MCP servers built both as production tools and as teaching material, Obsidian sync with markmap visualization, real estate platform integration (OAuth 2.1, Streamable HTTP), web article summarization with Playwright. Each one demonstrates a different transport, auth model, or tool design pattern.
I run training programs across very different student profiles: engineers, mathematicians, business professionals and have developed a consistent pedagogical philosophy: problem before solution, never introduce jargon before grounding it, intentional friction as a learning mechanism. Drawing on the Greek concept of aletheia and Socratic dialogue: knowledge surfaces through honest questioning, not through being handed answers.
Active courses include Ethical Hacking, Java (Spring Boot), Python, Machine Learning for cybersecurity, and MCP / Agent infrastructure.
sinensia.com My other github account: https://github.com/SINENSIA
For commercial licensing inquiries on Sinensia open-source projects, professional training engagements, or DevSecOps consulting — reach out via email.