From zero to a RAG system: successes and failures
An engineer spun up an internal chat with a localLLaMAmodel viaOllama, a PythonFlaskAPI, and aStreamlitfrontend. They moved off in-memoryLlamaIndexto batch ingestion intoChromaDB(SQLite). Checkpoints and tolerant parsing went in to stop RAM disasters. Indexing produced 738,470 vectors (~54 GB). They.. read more









