Llm

When format constraints hurt LLMs: the same structured-output techniques often lower scores on reasoning tasks and raise them on discrete classification—from agent pipelines to benchmark evaluation.

Semantic query engines treat foundation-model filter, join, classify, map, and rank as first-class operators—logical and physical plans, cost–quality tradeoffs, SemBench workloads, and how they differ from script-style RAG and vector search alone.

Judge-time compute: stacking structured, composable weak-model calls at evaluation time instead of assuming one expensive judge pass is enough—Verdict, agreement metrics, and production guardrails, with evidence boundaries called out.

Compound AI: When a single LLM call is not enough—multiple model calls, retrievers, tools, and business logic as a graph; structured output, specialist pipelines, inference stacks, and deployment granularity from a Weaviate podcast with Baseten’s Philip Kiely.

Agentic topic modeling: modular embedding pipelines, LLM-maintained topic tables, and human-in-the-loop granularity—engineering trade-offs between BERTopic, TopicGPT, and retrieval-scale deployment.