Girrasol Labs LLC

Independent AI Research Lab.

We conduct applied research in artificial intelligence — exploring agentic systems, retrieval-augmented generation, and small language model fine-tuning from our dedicated on-premises GPU laboratory.

What We Study

Research Areas

Agentic Systems

Investigating autonomous multi-step reasoning architectures that can plan, execute, and self-correct across complex task domains.

Retrieval-Augmented Generation

Developing novel retrieval pipelines and grounding techniques that improve factual accuracy and reduce hallucination in language model outputs.

Small Language Model Fine-Tuning

Researching efficient adaptation methods for compact models — enabling domain-specific performance without the cost and latency of large-scale inference.

The Lab

On-Premises Research Infrastructure

Our research is powered by a dedicated, on-premises multi-GPU laboratory. This gives us full control over the training and evaluation pipeline — enabling rapid experimentation, model fine-tuning, and reproducible results without reliance on third-party cloud compute.

Multi-GPU Compute

Dedicated on-premises hardware for training and inference

SLM Fine-Tuning

LoRA and QLoRA adaptation pipelines for compact models

Data Privacy

All data stays on-site — zero third-party exposure

Rapid Iteration

Full control over the stack from data ingestion to evaluation

Get in Touch

Contact Us

Interested in our research or exploring a collaboration? Drop us a line and we'll get back to you.