↻AI engineering

Walk through RAG, MCP, agent loops, evals, and token streaming as working flows. — 28 entries

···

Agent loop에이전트 루프

A model repeatedly chooses a tool, observes its result, and decides what to do next toward a goal.

···

Prompt engineering프롬프트 설계

State the goal, constraints, and examples so a model can respond in the intended format.

···

Context window문맥 창

The token budget available for a model’s input and output in one interaction.

···

Tokenization토큰화

Split input text into the token pieces a model processes.

···

Embeddings임베딩

Represent data such as text as numeric vectors that can be compared by similarity.

···

Semantic vector retrieval의미 벡터 검색

Embed queries and documents to retrieve material that is close in meaning.

···

Retrieval-augmented generation검색 증강 생성

Retrieve relevant material first, then give it to a model as context for its answer.

···

Document chunking문서 청킹

Split long documents into smaller passages for retrieval and model context.

···

Reranking재순위화

Rescore fast first-pass retrieval results with a more precise relevance model.

···

Hallucination근거 없는 생성

A model can produce plausible claims that are unsupported or conflict with evidence.

···

Temperature온도

A sampling setting that sharpens or flattens the probability distribution for the next token.

···

Guardrails가드레일

Place checks and limits around a model to handle disallowed inputs and outputs.

···

LLM evaluationLLM 평가

Compare model or prompt quality on representative tasks using explicit criteria.

···

Model routing모델 라우팅

Select different models or processing paths according to request characteristics.

···

Structured outputs구조화 출력

Receive model output in a defined shape with named fields and types.

···

Model Context Protocol모델 컨텍스트 프로토콜

A protocol for AI applications to discover and use external tools and resources.

···

Agent-to-Agent Protocol에이전트 간 프로토콜

A protocol for agents to discover one another and exchange tasks, status, and results.

···

Responses API

An API that represents a model response as output items such as messages and tool calls.

···

Agents SDK에이전트 SDK

A development toolkit for assembling agent runs, tools, handoffs, and traces in code.

···

Agent handoff에이전트 인계

Transfer control and relevant context from one agent to a specialist.

···

Prompt caching프롬프트 캐싱

Reuse repeated prompt prefixes to reduce processing work on later requests.

···

Top-p sampling누적 확률 샘플링

Sample from the smallest leading set of tokens whose cumulative probability reaches a threshold.

···

Fine-tuning미세 조정

Continue training a base model on examples to adapt its output behavior to a task.

···

Agents API

An API for managed agent execution environments and durable long-running sessions.

···

Built-in web search tool내장 웹 검색 도구

Let a model search the current web and attach sources to its answer.

···

Computer use agent컴퓨터 사용 에이전트

An agent that reads a screen and uses clicks and typing to work through a UI.

···

Agent code execution에이전트 코드 실행

Run agent-generated code in an isolated environment and feed the result back into the task.

···

Realtime voice agent실시간 음성 에이전트

Stream speech input and output so an agent can respond and handle interruptions.

Open as page ↗