sinanozel

Lightweight AI agent framework: run locally with Docker, deploy at scale on Kubernetes

Container to upscale images. Works as CLI, an HTTP server and an MCP tool. CUDA.

Cloud agnostic k3s deployment / teardown / clean up tool. https://github.com/sinan-ozel/k3s-anywhere

Containerized models served with llama optimized for 12GB VRAM NVidia GPUs using CUDA

A JupyterLab instance for image generation, with popular libraries and models that work in 12G VRAM

Lightweight AI agent framework: run locally with Docker, deploy at scale on Kubernetes

A JupyterLab container with popular libraries for generative AI for images, based on CUDA

Containerized models served with llama optimized for 6GB VRAM NVidia GPUs using CUDA

An MCP server containing a few financial tools.

A k8s-ready JupyterHub environment for multi-user data science. Uses GitHub to authenticate.

Containerized models, tested and ready for production, focused on smaller LLMs.

An MCP server for generating handouts and assets for Fantasy Role Playing games

This is an MCP server mainly for testing purposes. It has a few tools for looking up Eberron info

Model servers based on Ollama with pre-downloaded models, allowing version freeze and quick startup.

The JupyterLab environment that I use for working with data.

A server to generate spell cards for D&D 5e

Model servers based on Ollama with pre-downloaded models, allowing version freeze and quick startup.

A custom JupyterLab container for AI-related work, focusing on using GPU memory.

Model servers based on Ollama with pre-downloaded models, allowing version freeze and quick startup.

Model servers based on Ollama with pre-downloaded models, allowing version freeze and quick startup.

Model servers based on Ollama with pre-downloaded models, allowing version freeze and quick startup.

A JupyterLab instance for image generation, with popular libraries and models that work in 32G VRAM

Model servers based on Ollama with pre-downloaded models, allowing version freeze and quick startup.