ai-pipeline-orchestration – skill Agent & Tự động hóa cho agent AI

AI Pipeline Orchestration

Build reliable, observable AI workflows — from document ingestion to batch inference to model training pipelines.

When to Use This Skill

Use this skill when:

• Scheduling recurring RAG document ingestion and re-indexing
• Orchestrating multi-step batch LLM processing workflows
• Running nightly model evaluation and fine-tuning jobs
• Building ETL pipelines that feed into AI models
• Managing dependencies between data preparation and model serving

Execution contract

These are application design fragments: adapters, datasets, model files, secrets, vector schemas and deployment handlers are not bundled. Pin supported framework/provider versions, select the actual target and permissions, and authorize schedules, provider charges and writes separately. Use synthetic fixtures first; preserve private data, validate fetched sources, and persist versioned checkpoints to avoid repeating side effects.

Tool Selection

| Tool | Best For | Complexity | GPU Jobs |

|------|----------|------------|----------|

| Prefect | Modern Python-first; easy to adopt | Low | Good |

| Airflow | Complex DAGs; large teams; existing usage | High | Good |

| Dagster | Asset-centric; strong data lineage | Medium | Excellent |

| Temporal | Long-running workflows; reliability-first | Medium | Good |

Prefect — Quick Start

pip install prefect prefect-kubernetes

# Start Prefect server (or use Prefect Cloud)
prefect server start

# In another terminal
prefect worker start --pool default-agent-pool

Prefect: RAG ingestion design sketch (not executable)

from prefect import flow, task, get_run_logger
from prefect.tasks import task_input_hash
from datetime import timedelta
import hashlib

@task(cache_key_fn=task_input_hash, cache_expiration=timedelta(hours=24))
def fetch_documents(source_url: str) -> list[dict]:
    """Fetch documents from source; cached to avoid re-fetching."""
    logger = get_run_logger()
    logger.info(f"Fetching from {source_url}")
    # ... fetch logic
    return documents

@task(retries=3, retry_delay_seconds=30)
def chunk_and_embed(documents: list[dict]) -> list[dict]:
    """Chunk documents and generate embeddings with retry on failure."""
    from sentence_transformers import SentenceTransformer
    model = SentenceTransformer("BAAI/bge-large-en-v1.5")
    chunks = []
    for doc in documents:
        doc_chunks = chunk_text(doc["content"])
        embeddings = model.encode(doc_chunks, batch_size=64)
        for chunk, emb in zip(doc_chunks, embeddings):
            chunks.append({"text": chunk, "embedding": emb.tolist(),
                           "source": doc["url"], "doc_hash": doc["hash"]})
    return chunks

@task(retries=2)
def upsert_to_vector_store(chunks: list[dict]) -> int:
    """Upsert embeddings to Qdrant, skip unchanged documents."""
    from qdrant_client import QdrantClient
    client = QdrantClient("http://qdrant:6333")
    client.upsert(collection_name="knowledge-base", points=[...])
    return len(chunks)

@flow(name="rag-ingestion", log_prints=True)
def rag_ingestion_pipeline(sources: list[str]):
    """Full RAG ingestion flow — runs daily."""
    logger = get_run_logger()
    total = 0
    for source in sources:
        docs = fetch_documents(source)
        chunks = chunk_and_embed(docs)
        count = upsert_to_vector_store(chunks)
        total += count
        logger.info(f"Ingested {count} chunks from {source}")
    logger.info(f"Pipeline complete: {total} total chunks indexed")

if __name__ == "__main__":
    rag_ingestion_pipeline.serve(
        name="daily-rag-ingestion",
        cron="0 2 * * *",          # 2 AM daily
        parameters={"sources": ["https://docs.myapp.com", "https://api.myapp.com/kb"]},
    )

Prefect: Batch LLM Inference Pipeline

from prefect import flow, task
from prefect.concurrency.asyncio import concurrency
import asyncio
from openai import AsyncOpenAI

@task(retries=0)  # A batch retry can repeat successful billable calls
async def process_batch(items: list[dict], model: str) -> list[dict]:
    """Process a batch of items through LLM with rate limit protection."""
    client = AsyncOpenAI(timeout=60.0, max_retries=0)
    async with concurrency("openai-api", occupy=len(items), timeout_seconds=30, strict=True):  # Concurrent slots, not requests per minute
        tasks = [
            client.chat.completions.create(
                model=model,
                messages=[{"role": "user", "content": item["prompt"]}],
                max_tokens=256,
            )
            for item in items
        ]
        responses = await asyncio.gather(*tasks, return_exceptions=True)

    await client.close()
    results = []
    for item, response in zip(items, responses):
        if isinstance(response, Exception):
            results.append({**item, "error": type(response).__name__, "output": None})
        else:
            results.append({**item, "output": response.choices[0].message.content})
    return results

@flow(name="batch-llm-inference")
async def batch_inference_flow(input_file: str, output_file: str, model: str, batch_size: int = 10):
    import json
    if not 1 <= batch_size <= 10:
        raise ValueError("batch_size must be between 1 and 10")
    with open(input_file, encoding="utf-8") as source:
        items = []
        for line in source:
            if len(line.encode("utf-8")) > 32768 or len(items) >= 10000:
                raise ValueError("Input exceeds reviewed fixture limits")
            item = json.loads(line)
            if not isinstance(item, dict) or not isinstance(item.get("prompt"), str):
                raise ValueError("Invalid input item")
            items.append(item)
    batches = [items[i:i+batch_size] for i in range(0, len(items), batch_size)]

    all_results = []
    for batch in batches:
        results = await process_batch(batch, model)
        all_results.extend(results)

    # A full implementation reserves the output/checkpoint before any billed calls;
    # late file creation can otherwise fail after the whole batch has been charged.
    with open(output_file, "x", encoding="utf-8") as f:
        for result in all_results:
            f.write(json.dumps(result) + "\n")
    failed = sum(result.get("output") is None for result in all_results)
    return {"total": len(all_results), "failed_or_empty": failed, "completed": len(all_results) - failed}

The async batch fragment requires a pre-created limit large enough for each batch; it fails if absent. Store outputs in an authorized private directory, handle existing/partial files explicitly, and reconcile every provider call on timeout before retrying. Close the client in a finally block in a full implementation. Semaphore slots do not alone enforce token/RPM or currency budgets.

Airflow: training DAG design sketch (not executable)

from airflow.decorators import dag, task
from airflow.providers.cncf.kubernetes.operators.pod import KubernetesPodOperator
from datetime import datetime
from kubernetes.client import models as k8s

@dag(
    dag_id="llm_fine_tuning",
    schedule="@weekly",
    start_date=datetime(2025, 1, 1),
    catchup=False,
    tags=["ai", "training"],
)
def llm_fine_tuning_dag():

    @task
    def prepare_dataset() -> str:
        """Download and preprocess training data."""
        # ... data prep logic
        return "s3://my-bucket/training-data/2025-03-01/"

    train = KubernetesPodOperator(
        task_id="train_model",
        name="llm-training-job",
        namespace="ml",
        image="nvcr.io/nvidia/pytorch:24.05-py3",
        cmds=["accelerate", "launch", "-m", "axolotl.cli.train", "/config/config.yaml"],
        container_resources=k8s.V1ResourceRequirements(
            limits={"nvidia.com/gpu": "4", "memory": "320Gi"},
            requests={"nvidia.com/gpu": "4"},
        ),
        node_selector={"nvidia.com/gpu.product": "A100-SXM4-80GB"},
        volumes=[...],
        volume_mounts=[...],
        get_logs=True,
        is_delete_operator_pod=True,
    )

    @task
    def evaluate_model(dataset_path: str) -> dict:
        """Run evals; fail pipeline if quality drops."""
        metrics = run_evals()
        if metrics["accuracy"] < 0.85:
            raise ValueError(f"Model quality too low: {metrics}")
        return metrics

    @task
    def deploy_model(metrics: dict):
        """Push merged model to HF Hub and update vLLM config."""
        update_serving_config(new_model="org/fine-tuned-v2")

    dataset = prepare_dataset()
    train.set_upstream(dataset)
    eval_result = evaluate_model(dataset)
    eval_result.set_upstream(train)
    deploy_model(eval_result)

llm_fine_tuning_dag()

Dagster: asset pipeline design sketch (not executable)

from dagster import asset, AssetExecutionContext, define_asset_job, ScheduleDefinition

@asset(description="Raw documents fetched from knowledge sources")
def raw_documents(context: AssetEx

… (nội dung đầy đủ có trong gói tải về)