Lets Discuss AI
So I have one thing to disclose to everyone that is here, Yes, I do use LLMs and other coding assistants and security tools that work with AI(LLMs). Yeah bummer right? If yes you can leave this site but if not then fine continue.
About my usage, it mostly depends on menial labor like skimming through texts, logs and low priority work.
Not to be confused with major LLM providers, I only use local LLMs to maintain privacy. I have this simple gpu cluster that I use to host 2 models. I had tried most of their options, these are great and powerful but I don’t need them tbh.
I primarily use qwen3.5 since that is the more reliable one out of the two but I use llama 3.1 for repetitive tasks that are not that complicated.
I use llama to generate me resume using a particular flow where it gets all the details related to my professional experience and stuff like that and then it can create an upto date resume for me for that job profile or being ATS friendly. It is kind of a normal chatllm which I talk to using a simple streamlit app.
For the qwen model I use it for doing projects and analysing logs and sometime generating text that I am too lazy to write about, also sometimes I use it with opencode with oh-my-openagent for experimenting and log analysis that require more repetitive tasks.
That is it, I don’t use any of the fancy “Agentic” system such as openclaw or any proprietary shit, I only support open models available at hugging face that can be locally run on a mid size gpu node.
That is it for this blog, I hope I made myself clear, I dont ship slop to any of my public repo, I mostly use these LLM/Agents to analyse security related stuff during work or outside work.