KiwiFARS: Self-hosted LLM Research Infrastructure on HPC
What it is
KiwiFARS is a self-hosted infrastructure for LLM research on HPC clusters, specifically designed for SLURM environments. It combines private vLLM inference for large models like Qwen and DeepSeek with a 9-stage automated experiment pipeline and a RAG-based literature search system.
The problem it solves
Lack of integrated, self-hosted tools for LLM inference, experiment orchestration, and literature retrieval within HPC/SLURM environments.
What you could use it for
- Running private, local LLM inference on HPC clusters without external API costs
- Automating multi-stage scientific experiments with SLURM job orchestration
- Performing literature research using a local RAG system with FAISS vector search
- Generating daily automated digests of new arXiv papers with LLM summaries
Read the source ↗Back to the library →
Summary written by an AI model from the source; check the source before you rely on it.