New York Site views —Page views —
← All series

Language-model engineering

Practical guides to fine-tuning, deployment, tool use and retrieval.

Reading order

  1. Choosing between LoRA, QLoRA and full fine-tuning
  2. FP16, INT8 or 4-bit? Throughput, accuracy and memory in deployment
  3. Serving language models under load: monitoring, rate limits and recovery
  4. Function calling, tool use and agent frameworks
  5. How RAG fills knowledge gaps: retrieval, chunking, reranking and generation