Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/vrashmanyu605-eng/Langchain_Interview_Multi_Agents_Flow/llms.txt

Use this file to discover all available pages before exploring further.

The JD analysis agent runs in parallel with the resume parser — both can be called as soon as their respective raw inputs are available. It receives the raw job description text and asks the LLM to produce a structured breakdown of what the role requires. The resulting jd_analysis field is consumed by the matching agent, both interview question generators, the candidate research agent, and ultimately the evaluation agent.

​
Source code

from llm import llm


def jd_analysis_agent(state):

    jd_text = state["jd_text"]

    response = llm.invoke(
        f"""
        You are a JD Analysis Agent.

        Analyze this job description deeply.

        Job Description:
        {jd_text}

        Extract:
        - role_title
        - seniority_level
        - required_skills
        - preferred_skills
        - responsibilities
        - leadership_requirements
        - hiring_priorities

        Return structured JSON only.
        """
    )

    return {
        "jd_analysis": response.content
    }

​
Input

​
jd_text
string
required
The raw job description text provided as user input. This is typically copied directly from a job posting.

​
Output

​
jd_analysis
string
A JSON string containing the structured analysis of the job description. Consumed by the matching agent, candidate research agent, all three interview agents, and the evaluation agent.
The JSON contains the following fields:
FieldDescription
role_titleThe name of the role being hired for
seniority_levelExpected level (e.g., Junior, Senior, Staff, Principal)
required_skillsSkills listed as mandatory in the JD
preferred_skillsSkills listed as nice-to-have
responsibilitiesKey duties and accountabilities described in the JD
leadership_requirementsAny people management or mentoring expectations
hiring_prioritiesThe LLM’s inference of what the hiring team values most

​
LLM prompt

The prompt instructs the LLM to act as a JD Analysis Agent and perform a deep analysis of the provided jd_text. It requests all seven fields and asks for structured JSON only, keeping the output machine-readable for downstream agents.

​
Example output structure

{
  "role_title": "Senior Backend Engineer",
  "seniority_level": "Senior",
  "required_skills": ["Python", "REST APIs", "PostgreSQL", "AWS"],
  "preferred_skills": ["Kubernetes", "Redis", "GraphQL"],
  "responsibilities": ["Design scalable microservices", "Lead technical reviews"],
  "leadership_requirements": "Mentor junior engineers",
  "hiring_priorities": ["System design", "Python expertise", "AWS experience"]
}

​
State diagram

Requires: jd_text
    ↓
jd_analysis_agent
    ↓
Provides: jd_analysis

Build docs developers (and LLMs) love

Get started for freeTalk to us