Practical LangChain: From Prompts to Production

Building useful GenAI features means more than calling an LLM — you need structure, repeatability, and integration with data sources. That’s where LangChain shines.

Prompt Templates Done Right

Prompt templates give you reusable, parameterized text for LLM inputs:

from langchain.prompts import PromptTemplate

template = PromptTemplate(
    input_variables=["query"],
    template="Answer the following: {query}",
)

Using Retrievers to Add Context

from langchain.vectorstores import Chroma
from langchain.embeddings import OpenAIEmbeddings

embeddings = OpenAIEmbeddings()
vectorstore = Chroma.from_documents(docs, embeddings)

Stateful Agents

from langchain.agents import create_agent

agent = create_agent(
    model="gpt-4o",
    tools=[search_tool],
    system_prompt="You are a smart assistant",
)

Example: Summarizer Pipeline

from langchain.chains import RetrievalQAWithSourcesChain

qa_chain = RetrievalQAWithSourcesChain.from_llm(
    llm=OpenAI(temperature=0),
    retriever=vectorstore.as_retriever()
)

print(qa_chain("Summarize the key points from our docs"))

Final Thoughts

LangChain abstracts common generative AI patterns — prompt orchestration, RAG, agents, memory — into cohesive blocks.