Step 4: The Job Analyzer

# state.py
# ==========================================
# Shared state for the job application pipeline
# ==========================================
# Every node in the graph reads from and writes to this state.
# Each field represents one stage of the pipeline's output.
 
from typing import TypedDict
 
 
class JobApplicationState(TypedDict, total=False):
    # --- Inputs ---
    job_posting: str            # Raw job posting text
    resume: str                 # User's original resume (Markdown)
 
    # --- Job Analyzer output ---
    company_name: str           # Extracted company name
    job_title: str              # Extracted job title
    requirements: str           # Structured requirements and qualifications
    key_skills: str             # Comma-separated list of required skills
    culture_notes: str          # Notes about company culture and values
 
    # --- Resume Tailor output ---
    tailored_resume: str        # Rewritten resume targeting this role
 
    # --- Cover Letter Writer output ---
    cover_letter: str           # Targeted cover letter
 
    # --- Application Reviewer output ---
    review: str                 # Quality review with scores and suggestions
    review_passed: bool         # Whether the application meets quality bar

# agents/analyzer.py
# ==========================================
# Job Analyzer — extracts structured info from job postings
# ==========================================
 
from state import JobApplicationState
from utils import get_llm
 
 
ANALYZER_PROMPT = """You are an expert job posting analyst. Your task is to
extract structured information from a job posting.
 
Analyze the following job posting and extract:
 
1. **Company Name**: The hiring company
2. **Job Title**: The exact role title
3. **Requirements**: A structured list of:
   - Required years of experience
   - Required technical skills
   - Required soft skills
   - Required education or certifications
   - Nice-to-have qualifications
4. **Key Skills**: A comma-separated list of the most important technical
   skills mentioned (these will be used to tailor a resume)
5. **Culture Notes**: Any hints about company culture, values, work style,
   or what kind of person would thrive in this role
 
Format your response exactly like this:
 
COMPANY: [company name]
TITLE: [job title]
 
REQUIREMENTS:
[structured bullet list]
 
KEY SKILLS: [skill1, skill2, skill3, ...]
 
CULTURE NOTES:
[observations about culture and values]
 
---
 
Job Posting:
{job_posting}"""
 
 
def analyze_job(state: JobApplicationState) -> dict:
    """Extract structured information from the job posting."""
    llm = get_llm()
 
    prompt = ANALYZER_PROMPT.format(job_posting=state["job_posting"])
    response = llm.invoke(prompt)
    content = response.content
 
    # Parse the structured response
    result = {
        "requirements": content,  # Full analysis for downstream agents
        "key_skills": "",
        "company_name": "",
        "job_title": "",
        "culture_notes": "",
    }
 
    # Extract individual fields from the formatted response
    for line in content.split("\n"):
        line = line.strip()
        if line.startswith("COMPANY:"):
            result["company_name"] = line.replace("COMPANY:", "").strip()
        elif line.startswith("TITLE:"):
            result["job_title"] = line.replace("TITLE:", "").strip()
        elif line.startswith("KEY SKILLS:"):
            result["key_skills"] = line.replace("KEY SKILLS:", "").strip()
 
    # Extract culture notes (everything after "CULTURE NOTES:")
    if "CULTURE NOTES:" in content:
        result["culture_notes"] = content.split("CULTURE NOTES:")[-1].strip()
 
    return result

# test_analyzer.py
# ==========================================
# Test the Job Analyzer in isolation
# ==========================================
 
from agents.analyzer import analyze_job
 
# A realistic job posting for testing
test_posting = """
Senior Python Developer — DataFlow Inc.
 
About Us:
DataFlow is a fast-growing data infrastructure startup backed by top-tier VCs.
We value ownership, clear communication, and shipping fast. Our team is remote-first
and we believe in async collaboration.
 
The Role:
We are looking for a Senior Python Developer to join our platform team.
You will design and build the core data pipeline services that power our product.
 
Requirements:
- 5+ years of professional Python experience
- Strong experience with FastAPI or Django
- Experience with distributed systems and message queues (Kafka, RabbitMQ)
- Proficiency in PostgreSQL and Redis
- Familiarity with Docker and Kubernetes
- Experience with CI/CD pipelines
- Strong communication skills and ability to work independently
 
Nice to Have:
- Experience with Apache Spark or Flink
- Contributions to open-source projects
- Experience mentoring junior developers
 
Benefits:
- Competitive salary ($180k-$220k)
- Remote-first culture
- Equity package
- Unlimited PTO
"""
 
result = analyze_job({"job_posting": test_posting, "resume": ""})
 
print(f"Company: {result['company_name']}")
print(f"Title: {result['job_title']}")
print(f"Key Skills: {result['key_skills']}")
print(f"\nFull Analysis:\n{result['requirements']}")

python test_analyzer.py

rm test_analyzer.py