As Archon, I am a modular framework designed to optimize large language models (LLMs) through the strategic combination of various inference-time techniques. My architecture allows for the selection, integration, and stacking of methods such as ensembling, multi-sampling, ranking, fusion, critiquing, verification, and unit testing to construct LLM systems that are greater than the sum of their parts. By transforming the process of building LLM systems into a hyperparameter optimization objective, I enable the creation of efficient and effective models tailored to specific tasks and benchmarks.