Step 7: The Application Reviewer

# agents/reviewer.py
# ==========================================
# Application Reviewer — quality gate for the pipeline
# ==========================================
 
from state import JobApplicationState
from utils import get_llm
 
 
REVIEWER_PROMPT = """You are a senior hiring manager and career coach reviewing
a job application package. Your job is to evaluate the quality, consistency,
and completeness of the application materials.
 
## Job Requirements
Company: {company_name}
Title: {job_title}
 
Key Skills Required: {key_skills}
 
Full Requirements:
{requirements}
 
## Tailored Resume
{tailored_resume}
 
## Cover Letter
{cover_letter}
 
## Your Review Criteria
 
Score each category from 1-5 and provide specific feedback:
 
1. **Keyword Match (1-5)**: Do the resume and cover letter include the
   key skills from the job posting? List any missing keywords.
 
2. **Relevance (1-5)**: Is the experience framed in terms relevant to
   this specific role? Or does it read like a generic resume?
 
3. **Consistency (1-5)**: Do the resume and cover letter tell a coherent
   story? Are there contradictions or mismatches?
 
4. **Conciseness (1-5)**: Is the writing tight and professional? Or is
   it verbose and padded?
 
5. **Authenticity (1-5)**: Does the application sound genuine? Flag
   anything that seems fabricated or exaggerated.
 
## Output Format
 
SCORES:
- Keyword Match: [1-5]
- Relevance: [1-5]
- Consistency: [1-5]
- Conciseness: [1-5]
- Authenticity: [1-5]
- Overall: [1-5]
 
STRENGTHS:
[What works well — be specific]
 
IMPROVEMENTS:
[What should be changed — be specific and actionable]
 
VERDICT: [PASS or NEEDS_REVISION]
 
A PASS requires an overall score of 4 or higher. Be honest but constructive."""
 
 
def review_application(state: JobApplicationState) -> dict:
    """Review the complete application package."""
    llm = get_llm()
 
    prompt = REVIEWER_PROMPT.format(
        company_name=state.get("company_name", "Unknown"),
        job_title=state.get("job_title", "Unknown"),
        key_skills=state.get("key_skills", ""),
        requirements=state.get("requirements", ""),
        tailored_resume=state.get("tailored_resume", "No resume provided"),
        cover_letter=state.get("cover_letter", "No cover letter provided"),
    )
 
    response = llm.invoke(prompt)
    content = response.content
 
    # Parse the verdict
    review_passed = "VERDICT: PASS" in content.upper()
 
    return {
        "review": content,
        "review_passed": review_passed,
    }

                    ┌─── review_passed=True ───► Save & Finish
Application        │
Reviewer ──────────┤
                    └─── review_passed=False ──► Revise (loop back)

# test_reviewer.py
# ==========================================
# Test the Application Reviewer with mock data
# ==========================================
 
from agents.reviewer import review_application
 
mock_state = {
    "company_name": "DataFlow Inc.",
    "job_title": "Senior Python Developer",
    "key_skills": "Python, FastAPI, PostgreSQL, Redis, Docker, Kubernetes, Kafka, CI/CD",
    "requirements": """
    - 5+ years Python experience
    - Strong FastAPI or Django experience
    - Distributed systems (Kafka, RabbitMQ)
    - PostgreSQL and Redis
    - Docker and Kubernetes
    """,
    "tailored_resume": """# Alex Chen
**Senior Python Developer** | San Francisco, CA
 
## Experience
### Software Engineer — Acme Corp (2022–Present)
- Architected high-throughput REST APIs with Python and FastAPI,
  serving 50k+ daily active users with PostgreSQL and Redis
- Led migration from monolith to microservices on AWS with Docker
  and Kubernetes, reducing deployment time by 60%
- Built CI/CD pipelines with GitHub Actions for automated testing
  and deployment across 12 microservices
 
## Skills
Python, FastAPI, PostgreSQL, Redis, Docker, Kubernetes, AWS, CI/CD, Git
""",
    "cover_letter": """Dear Hiring Manager,
 
DataFlow's mission to build next-generation data infrastructure caught my
attention — particularly your focus on high-throughput pipeline services.
 
In my current role at Acme Corp, I have spent the last two years doing
exactly this: designing FastAPI services that handle 50k+ daily users,
backed by PostgreSQL and Redis. I led our migration from a monolith to
a microservices architecture on Docker and Kubernetes, which maps directly
to the distributed systems work your team is tackling.
 
I thrive in remote-first, async environments — my team at Acme operates
across three time zones, and clear written communication is how we ship.
 
I would welcome the chance to discuss how my experience with Python-based
data services could accelerate DataFlow's platform roadmap.
 
Best regards,
Alex Chen""",
    "job_posting": "",
    "resume": "",
    "culture_notes": "",
}
 
result = review_application(mock_state)
print(result["review"])
print(f"\nPassed: {result['review_passed']}")

python test_reviewer.py

rm test_reviewer.py