LLM Ops

LLM-OPS -- IA de Producao

Overview

LLM Operations -- RAG, embeddings, vector databases, fine-tuning, prompt engineering avancado, custos de LLM, evals de qualidade e arquiteturas de IA para producao. Ativar para: implementar RAG, criar pipeline de embeddings, Pinecone/Chroma/pgvector, fine-tuning, prompt engineering, reducao de custos de LLM, evals, cache semantico, streaming, agents.

When to Use This Skill

• When you need specialized assistance with this domain

Do Not Use This Skill When

• The task is unrelated to llm ops
• A simpler, more specific tool can handle the request
• The user needs general-purpose assistance without domain expertise

How It Works

A diferenca entre um prototipo de IA e um produto de IA e operabilidade. LLM-Ops e a engenharia que torna IA confiavel, escalavel e economica.

Arquitetura Rag Completa

[Documentos] -> [Chunking] -> [Embeddings] -> [Vector DB] | [Query] -> [Embed query] -> [Semantic Search] -> [Top K chunks] | [LLM + Context] -> [Resposta]

Pipeline De Indexacao

from anthropic import Anthropic import chromadb

client = Anthropic()
chroma = chromadb.PersistentClient(path="./chroma_db")

def chunk_text(text, chunk_size=500, overlap=50):
    words = text.split()
    chunks = []
    for i in range(0, len(words), chunk_size - overlap):
        chunk = " ".join(words[i:i + chunk_size])
        if chunk: chunks.append(chunk)
    return chunks

def index_document(doc_id, content_text, metadata=None):
    chunks = chunk_text(content_text)
    ids = [f"{doc_id}_chunk_{i}" for i in range(len(chunks))]
    collection.upsert(ids=ids, documents=chunks)
    return len(chunks)

Pipeline De Query Com Rag

def rag_query(query, top_k=5, system=None): results = collection.query( query_texts=[query], n_results=top_k, include=["documents", "metadatas", "distances"]) context_parts = [] for doc, meta, dist in zip(results["documents"][0], results["metadatas"][0], results["distances"][0]): if dist < 1.5: src = meta.get("source", "doc") context_parts.append(f"[Fonte: {src}] {doc}") context = "

".join(context_parts) response = client.messages.create( model="claude-opus-4-20250805", max_tokens=1024, system=system or "Responda baseado no contexto.", messages=[{"role": "user", "content": f"Contexto: {context}

{query}"}]) return response.content[0].text

Escolha Do Vector Db

Pgvector

CREATE EXTENSION IF NOT EXISTS vector; CREATE TABLE knowledge_embeddings ( id UUID PRIMARY KEY DEFAULT gen_random_uuid(), content TEXT NOT NULL, embedding vector(1536), metadata JSONB, created_at TIMESTAMPTZ DEFAULT NOW() ); CREATE INDEX ON knowledge_embeddings USING ivfflat (embedding vector_cosine_ops) WITH (lists = 100); SELECT content, 1 - (embedding <=> QUERY_VECTOR) AS similarity FROM knowledge_embeddings ORDER BY similarity DESC LIMIT 5;

Estrutura De Prompt De Elite

Componentes do system prompt Auri:

• Identidade: Nome (Auri), Tom (Natural, caloroso, direto), Plataforma (Amazon Alexa)
• Regras: Maximo 3 paragrafos curtos, sem markdown, linguagem conversacional
• Capacidades: analise de negocios, conselho baseado em dados, criatividade
• Limitacoes: sem internet tempo real, sem transacoes financeiras
• Personalizacao: {user_name}, {user_preferences}, {relevant_history}

Chain-Of-Thought

def cot_analysis(problem: str) -> str: steps = [ "1. O que exatamente esta sendo pedido?", "2. Que informacoes sao criticas para resolver?", "3. Quais abordagens possiveis existem?", "4. Qual abordagem e melhor e por que?", "5. Quais riscos ou limitacoes existem?", ] prompt = f"Analise passo a passo:

PROBLEMA: {problem}

" prompt += " ".join(steps) + "

Resposta final (concisa, para voz):" return call_claude(prompt)

Cache Semantico

class SemanticCache: def init(self, similarity_threshold=0.95): self.threshold = similarity_threshold self.cache = {}

    def get_cached(self, query, embedding):
        for cached_emb, (response, _) in self.cache.items():
            if cosine_similarity(embedding, cached_emb) >= self.threshold:
                return response
        return None

    def set_cache(self, query, embedding, response):
        self.cache[tuple(embedding)] = (response, query)

Estimativa De Custos Claude

PRICING = { "claude-opus-4-20250805": {"input": 15.00, "output": 75.00}, "claude-sonnet-4-5": {"input": 3.00, "output": 15.00}, "claude-haiku-3-5": {"input": 0.80, "output": 4.00}, }

def estimate_monthly_cost(model, avg_input, avg_output, req_per_day):
    p = PRICING[model]
    daily = (avg_input + avg_output) * req_per_day / 1e6
    monthly = daily * p["input"] * 30
    return {"model": model, "monthly_cost": "USD %.2f" % monthly}

Framework De Avaliacao

from anthropic import Anthropic client = Anthropic()

def evaluate_response(question, expected, actual, criteria):
    criteria_text = "

".join(f"- {c}" for c in criteria) eval_prompt = ( f"Avalie a resposta do assistente de IA.

" f"PERGUNTA: {question} RESPOSTA ESPERADA: {expected} " f"RESPOSTA ATUAL: {actual}

Criterios: {criteria_text}

" "Nota 0-10 e justificativa para cada criterio. Formato JSON." ) response = client.messages.create( model="claude-haiku-3-5", max_tokens=1024, messages=[{"role": "user", "content": eval_prompt}] ) import json return json.loads(response.content[0].text)

AURI_EVALS = [
    {
        "question": "Quais sao os principais riscos de abrir startup agora?",
        "criteria": ["precisao_factual", "relevancia", "clareza_para_voz"]
    },
]

6. Comandos

Best Practices

• Provide clear, specific context about your project and requirements
• Review all suggestions before applying them to production code
• Combine with other complementary skills for comprehensive analysis

Common Pitfalls

• Using this skill for tasks outside its domain expertise
• Applying recommendations without understanding your specific context
• Not providing enough project context for accurate analysis

Limitations

• Use this skill only when the task clearly matches the scope described above.
• Do not treat the output as a substitute for environment-specific validation, testing, or expert review.
• Stop and ask for clarification if required inputs, permissions, safety boundaries, or success criteria are missing.