MicheleLacorte
Show HN: Meddle – AI-powered IIoT platform for small manufacturers
Show HN: Meddle – Integrate, collect, manage and analyze realtime data
Hello HN, I’m Michele. I’m building Meddle because connecting hardware stuff is still way harder than it should be.Every factory I worked with had the same problem: robots, PLCs, CNCs, sensors, all speaking different protocols, all locked in silos. Getting data out meant custom scripts, fragile gateways, or expensive vendor lock-ins.What I built: Meddle is a data-driven IoT platform that...
Show HN: Bookmarks in one smart place with SaveTo
How often do we save links or texts that we then never retrieve because we forget where they are or simply cannot remember the name of the website?This happens to me often, so I decided to create SaveTo, a simple cross-platform app that allows you to save links that can then be retrieved via chat, without having to remember the exact name of a website or link library.I've just opened the...
Show HN: Kritikos – Ready to use Go back end for LLM-as-a-critique
https://github.com/michelelacorte-quinck/kritikos
Hello HN! I've been having this problem with VertexAI's RAG for some time now, basically if you ask Gemini (with RAG) for something and it doesn't have the answer, then it simply says it can't find it, without doing fallback to the general model.So, in documenting myself I came across LLM-as-a-critique (also LLM-as-judge), a methodology whereby a second LLM is used to evaluate the answers of...