AI / LangGraph LangChain Interview questions
What is the difference between sequential and parallel chains?
In a sequential chain, components run one after another: the output of step N becomes the input of step N+1. This is the default LCEL pipe behaviour — chain = step1 | step2 | step3 means step2 cannot start until step1 finishes.
In a parallel chain, multiple branches run concurrently on the same input, and their results are merged into a single dict. LangChain implements this with RunnableParallel:
from langchain_core.runnables import RunnableParallel, RunnablePassthrough from langchain_openai import ChatOpenAI from langchain_core.output_parsers import StrOutputParser llm = ChatOpenAI() parallel_chain = RunnableParallel( summary=summary_prompt | llm | StrOutputParser(), sentiment=sentiment_prompt | llm | StrOutputParser(), keywords=keywords_prompt | llm | StrOutputParser(), ) # Runs all three LLM calls concurrently, then returns: # {"summary": "...", "sentiment": "...", "keywords": "..."} result = parallel_chain.invoke({"text": "LangChain is amazing..."})
Use sequential chains when each step depends on the previous result. Use parallel chains when steps are independent of each other — this reduces wall-clock time to the slowest branch's latency rather than the sum of all branches.
What class in LangChain runs multiple branches concurrently on the same input?
When does using RunnableParallel reduce latency compared to sequential chaining?
More Related questions...
