Salvador Guzman

I am interested in open-source AI infrastructure, local LLM workflows, retrieval-augmented generation, symbolic computation, document processing, speech and text pipelines, and tools that connect machine learning with serious research work. My main focus is building practical systems around language models: document ingestion, entity extraction, knowledge graphs, semantic search, embeddings, OCR, text-to-speech, speech-to-text, and structured reasoning over large corpora. I am especially interested in using AI to process books, archives, technical documents, media, and multilingual material into searchable, computable knowledge. I also care about the bridge between machine learning and mathematics: symbolic AI, formal representations, combinatory logic, quantitative research, time series analysis, and model-assisted exploration of abstract structures. I prefer open-source, local-first, reproducible AI stacks. I like tools that can run on personal hardware, integrate cleanly with Linux, Rust, Python, Docker, and existing research workflows, and help individuals build serious intellectual infrastructure without depending entirely on closed platforms.