Recipe

Recipe: Resume Screener

Parse uploaded resumes, extract structured candidate data, and score each applicant against your job description using Meridian's LLM pipeline.

PDF + DOCX parsingJD keyword matchingBatch scoringCSV export

Overview

The Resume Screener recipe automates the initial candidate review process. Upload a batch of resumes alongside a job description, and Meridian extracts work history, skills, education, and certifications from each document. Every resume receives a relevance score (0–100) based on semantic alignment with the JD, plus a breakdown of matched and missing keywords. Results are presented in a sortable table with one-click CSV export for your ATS or spreadsheet.

How it works

Upload & Parse

Drop PDF or DOCX resumes into the pipeline. Meridian extracts raw text, identifies section boundaries, and structures the data into a canonical candidate profile.

Score Against JD

The LLM compares each profile against your job description, computing a composite relevance score from skill overlap, experience level, and keyword density.

Review & Export

Sort candidates by score, drill into individual match breakdowns, and export the full ranked list as CSV for downstream processing.

Configuration

Parameter	Type	Default	Description
score_threshold	number	60	Minimum relevance score to flag a candidate as "shortlisted"
max_resumes	number	100	Maximum resumes processed per batch run
include_raw_text	boolean	false	Include extracted raw text in the output payload
output_format	string	"csv"	Export format: csv or json

Sample output

Each resume produces a structured result object. Below is a trimmed example for a single candidate scored against a backend-engineer JD.

{
  "candidate_id": "res_0042",
  "filename": "jane_doe_resume.pdf",
  "extracted": {
    "name": "Jane Doe",
    "email": "jane@example.com",
    "phone": "+1-555-0142",
    "skills": ["Python", "PostgreSQL", "Docker", "FastAPI"],
    "years_experience": 5,
    "education": ["B.S. Computer Science"]
  },
  "scoring": {
    "relevance_score": 87,
    "skill_match_pct": 0.82,
    "matched_keywords": ["Python", "PostgreSQL", "Docker"],
    "missing_keywords": ["Kubernetes", "gRPC"],
    "experience_fit": "strong"
  }
}

API endpoint

Trigger the recipe programmatically via the Meridian API. Send a multipart upload with resumes and a job description string.

POST/api/v1/recipes/resume-screener/run

curl -X POST https://api.getnimbus.net/v1/recipes/resume-screener/run \
  -H "Authorization: Bearer $MERIDIAN_API_KEY" \
  -F "job_description=@jd.txt" \
  -F "resumes=@candidate1.pdf" \
  -F "resumes=@candidate2.docx" \
  -F "score_threshold=70" \
  -F "output_format=csv"

Ready to screen resumes at scale?

Add this recipe to your Meridian pipeline and start processing candidates in minutes.

Add to pipeline Read API docs