can you share a bit into the engineering into how this works? i'm hypothesizing: 1. chunk entire read.cv portfolio per designer 2. feed that as a param into a custom prompt 3. buttons function as queries into the LLM do you use langchain?
Had fun creating this little SMS inspired website template for Read.cv. It works great for all sorts of profiles and content, and is probably even better on mobile than it is desktop. Demo: 24okqxhg1.sites.cv/ Link to template: read.cv/sites/template…
what you described, i actually implemented a few weeks ago in my portfolio. it’s just a simple prompt for now - no RAG, no fine-tuning. uses meta/llama-3.1-8b-instruct-fast. sean.app/ai
lmao yes that’s why i was kinda hesistant to share it + fear of unexpected ai bills 😂 it’s just a proof of concept for now… the goal is kinda like AI Gilfoyle that’s trained on my messages, emails, and notes. like i want it to repond the way i would respond
This is like a portfolio site I started building at least twice, but much better and actually finished
works pretty well with with your content! joer-24okqxhg1.sites.cv
One of my favorite things about sites is I can just “try on” new designs. You guys have totally cured me of the painful part of “quarterly urge to redesign it” and I can just go right for the fun part.