2025

an archive of posts from this year

Dec 29, 2025 Similarity metrics for embeddings
Dec 25, 2025 Tokenizers: production economics cheat-sheet
Dec 19, 2025 The metric gap: bridging business outcomes and AI component optimization
Dec 15, 2025 Reflection vs evaluation: why the Agent-Critic pattern fails without separation of concerns
Dec 12, 2025 Vector search + hard filters in Elasticsearch: the hidden RAG bottleneck
Dec 09, 2025 Architecture design: a constraint-satisfaction approach
Dec 05, 2025 Classification with LLMs: getting accurate probabilities from structured output
Dec 02, 2025 Token optimization: three production patterns that reduce LLM costs by 70%
Nov 26, 2025 Hierarchical signal tuning: optimizing components before fusion
Nov 22, 2025 Jensen-Shannon divergence for meaningful clustering
Nov 19, 2025 Hybrid intent classification: the rationale for production-grade shallow-model-first architectures
Nov 15, 2025 Few-shot prompt ordering: the impact of example position
Nov 10, 2025 Temporal for LLM pipelines: durable execution starter pack
Nov 07, 2025 GraphRAG: beyond vector search for connecting the dots
Nov 04, 2025 Domain-driven design for AI systems: architectural patterns and production experience
Oct 30, 2025 Semantic prompt caching: when LLM-judge beats exact match
Oct 27, 2025 The reranking trap: when cross-encoders make things worse
Oct 24, 2025 Structured output engineering for production LLMs
Oct 20, 2025 The chunk size dilemma: identifying the optimal value in RAG systems
Oct 10, 2025 Mitigating positional bias in LLM-as-a-judge evaluation: the swapping technique
Oct 06, 2025 Hybrid retrieval with reciprocal rank fusion: solving the score normalization problem
Oct 02, 2025 LLM orchestration: a pragmatic guide to complexity
Sep 28, 2025 How Qdrant's scalar quantization cut our RAG latency by 3x
Sep 21, 2025 Why vision-language models ignore visual evidence (and how to fix it)
Sep 14, 2025 Our agents argued endlessly. Here's how a hybrid AI pattern tamed LLM chaos
Sep 09, 2025 Vision-language model pipeline debugging: lessons from visual monitoring
Sep 05, 2025 Machine learning metrics for undefined projects: 3 critical mistakes
Sep 02, 2025 Pragmatic LLM debugging: a survival guide to chaos