Federico Solutions, Myrko Federico Several AI agents working together without losing track One territory per agent, every request in a ticket, nothing irreversible on a relayed order: the rules I apply to my own AI agents, and the mistakes that gave rise to them. AI R&D Strategy & advisory Nov 16, 2026 Experiences
Federico Solutions, Myrko Federico Measure before you sell: the test bench for my AI models A detection model can only be sold if you know how many it misses. Here is how I test mine, and the questions to ask an AI supplier. AI R&D Nov 16, 2026 Experiences
Federico Solutions, Myrko Federico Local LLM without a graphics card: what holds, what crashes I ran a large language model on a mini-PC with no dedicated graphics card. What holds, what crashes, and why I do not write down a cause I have not established. AI R&D Systems & infrastructure Nov 16, 2026 Experiences
Federico Solutions, Myrko Federico A library and an AI that work without the internet An offline knowledge server with an assistant that cites its sources: why I started from Project NOMAD, and what it taught me. AI R&D Systems & infrastructure Nov 15, 2026 Experiences
Federico Solutions, Myrko Federico Connecting an AI assistant to Odoo through MCP My AI assistant has been reading and writing in Odoo since April 2026. What in this setup is a third-party tool, what I wrote myself, and what I take away. AI R&D Systems & infrastructure Nov 15, 2026 Experiences
Federico Solutions, Myrko Federico Self-hosted team chat to run your AI assistant Talking to your AI assistant from your phone without sending your conversations to a provider: why and how I connected it to a self-hosted team chat. AI Systems & infrastructure Nov 15, 2026 Experiences
Federico Solutions, Myrko Federico AI assistant memory: make it remember, on your server Your AI assistant forgets everything every evening. How I gave it a persistent memory, with a vector database and a graph, hosted on my own server. AI Systems & infrastructure Nov 15, 2026 Experiences
Federico Solutions, Myrko Federico AI agent for small business: my OpenClaw experience For several months, my AI assistant ran on my own server with OpenClaw. What worked, what got stuck, and the principles I keep whatever the tool. AI R&D Systems & infrastructure Nov 15, 2026 Experiences
Federico Solutions, Myrko Federico AI agent guardrails: before you hand over your tools After shadow AI comes the next step: letting AI act. The nine guardrails to put in place before giving an agent access to your e-mail, your ERP or your payments. AI Systems & infrastructure Nov 15, 2026 News
Federico Solutions, Myrko Federico Self-hosted AI: why a business should keep part in-house An online AI service is rented: the provider sets the model, the price and the retirement date. Here is how to split your uses between a local model and an external API. AI Systems & infrastructure Nov 15, 2026 News
Federico Solutions, Myrko Federico Run LLMs locally: big models on small machines Once compressed, an 8-billion-parameter model fits on a recent workstation. Memory, quality, hardware: what to know before you buy. AI Systems & infrastructure Nov 15, 2026 News
Federico Solutions, Myrko Federico Train an object detection model: licence first Before picking the best detection model, I read its licence. What I use for DataTerra, what I ruled out, and why you measure before you sell. AI R&D Systems & infrastructure Nov 15, 2026 Experiences