# Pemikiran & Cerita: AI/ML

> Berbagi pengalaman, pembelajaran, dan petualangan saya di dunia teknologi dan lainnya.

URL: https://www.matthewswong.com/id/blog/category/ai-ml/

- [Datadog Scout 1.0: Arti Investigasi AI SRE yang Lebih Murah](https://www.matthewswong.com/id/blog/datadog-scout-sre-model-bits-investigation-cost/)
- [LMCache Multi-Process Mode: Shared KV Cache untuk MoE Serving](https://www.matthewswong.com/id/blog/lmcache-multi-process-mode-moe-shared-kv-cache/)
- [vLLM Semantic Router v0.4: Panduan Routing Mixture-of-Models](https://www.matthewswong.com/id/blog/vllm-semantic-router-v0-4-mixture-of-models-routing/)
- [Panduan KServe LLMInferenceService: llm-d, WVA dan vLLM](https://www.matthewswong.com/id/blog/kserve-llminferenceservice-llm-d-production-guide/)
- [OCI Model Artifacts: Cold Start Model Kubernetes Lebih Cepat](https://www.matthewswong.com/id/blog/oci-model-artifacts-kserve-faster-model-cold-start/)
- [NVIDIA Dynamo di Kubernetes: Rust EPP dan Upgrade v1beta1](https://www.matthewswong.com/id/blog/nvidia-dynamo-1-5-rust-endpoint-picker-kubernetes/)
- [DeepSeek V4.1 Flash di vLLM vs SGLang: Arti Angkanya](https://www.matthewswong.com/id/blog/deepseek-v4-1-flash-serving-vllm-vs-sglang/)
- [SGLang Unified Radix Cache: Prefix Caching Model Hybrid](https://www.matthewswong.com/id/blog/sglang-unified-radix-cache-hybrid-model-prefix-caching/)
- [vllm-metal di Apple Silicon: Server vLLM Lokal untuk Mac](https://www.matthewswong.com/id/blog/vllm-metal-apple-silicon-local-llm-server/)
- [Release Notes vLLM 0.31: Preload, Model Runner V2, dan Flag yang Rusak Saat Upgrade](https://www.matthewswong.com/id/blog/vllm-0-31-fast-restart-preload-weight-cache/)
- [vLLM Bi-Directional KV Transfer untuk Agen Multi-Turn](https://www.matthewswong.com/id/blog/vllm-bidirectional-kv-transfer-multi-turn-agents/)
- [vLLM Disaggregated Prefill Decode: Panduan NixlConnector](https://www.matthewswong.com/id/blog/vllm-disaggregated-prefill-decode-serving-guide/)
- [vLLM NixlPushConnector: Push vs Pull KV Transfer dan TTFT](https://www.matthewswong.com/id/blog/vllm-nixl-push-connector-faster-ttft-disaggregated/)
- [Memilih Model Embedding: Benchmark Praktis](https://www.matthewswong.com/id/blog/embedding-model-selection-benchmark/)
- [Hybrid Search dan Reranking: RAG Retrieval yang Lebih Baik](https://www.matthewswong.com/id/blog/hybrid-search-reranking-rag/)
- [RAG vs Fine-Tuning: Framework Keputusan yang Jujur](https://www.matthewswong.com/id/blog/rag-vs-fine-tuning-decision/)
- [Semantic Search dengan pgvector: Lewati Vector Database Khusus](https://www.matthewswong.com/id/blog/pgvector-semantic-search-postgres/)
- [Membangun Aplikasi RAG LLM untuk Produksi: Arsitektur, Chunking, dan Evaluasi](https://www.matthewswong.com/id/blog/building-production-rag-llm-applications/)

---

- HTML version: https://www.matthewswong.com/id/blog/category/ai-ml/
- English: https://www.matthewswong.com/en/blog/category/ai-ml/
- Site index for agents: https://www.matthewswong.com/llms.txt
- Sitemap: https://www.matthewswong.com/sitemap.xml
- API description (OpenAPI): https://www.matthewswong.com/openapi.json