llm-fine-tuning-lora-qlora verified

Fine-tune large language models efficiently with LoRA/QLoRA and PEFT — rank/alpha/target-module choices, data prep, training config, and merge/export.

cross-encoder-reranking verified

Lift RAG/search precision with retrieve-then-rerank — bi-encoder first stage, cross-encoder second stage, rerank budgets, and measuring lift.

graphrag-knowledge-graphs verified

Build GraphRAG over your corpus — entity/relation extraction, entity resolution, community detection, and graph-enhanced retrieval.

hybrid-vector-keyword-search verified

Combine BM25 keyword and dense vector retrieval with reciprocal rank fusion (RRF) for search that catches both exact and semantic matches.

llm-as-judge-evaluation verified

Use an LLM as an automated evaluator for outputs — build the rubric, control position/length/self-preference bias, and validate against humans.

llm-context-window-management verified

Keep long agent and chat conversations inside the LLM context window — truncation, summarization, eviction, and two-layer memory.

llm-prompt-caching verified

Cut LLM API cost and latency with prompt caching — provider cache_control, prefix design, TTL, and hit-rate monitoring.

llm-quantization-gguf verified

Run LLMs locally with quantization — GGUF format, K-quant vs legacy levels, quality/size tradeoffs, and llama.cpp serving.

llm-red-teaming verified

Systematically attack and harden LLM apps — OWASP LLM Top 10, jailbreak/indirect-injection testing, and CI-integrated red teaming.

llm-semantic-caching verified

Cache semantically-similar LLM requests with embeddings to cut cost and latency — exact vs semantic layers, thresholds, and validation.