πŸ‘€
Available

Candidate Analyzer

Candidate Analyzer uses semantic embeddings and language models to score candidates across 5 weighted dimensions: technical skills (30%), work experience (25%), education (15%), languages (15%), and location (15%). The service analyzes structured or raw resumes, compares them against job descriptions, and generates a detailed report with recommendations. The matching algorithm goes beyond keywords to understand the real meaning of skills and experience.

EmbeddingsMatchingScoringHR

How it works

1

Send the resume and job description

Submit the candidate's resume (PDF, DOCX, or structured JSON) and the description of the role to fill.

2

Semantic vectorization

The resume and job description are converted into semantic vectors for comparison beyond simple keywords.

3

Multi-dimensional scoring

The candidate is evaluated across 5 weighted axes: skills (30%), experience (25%), education (15%), languages (15%), location (15%).

4

Report and recommendations

A detailed report is generated with the overall score, per-dimension scores, strengths, and points of attention.

API call example

analyze_candidate.py
import httpx

response = httpx.post(
    class="text-teal">"https:class="text-text-faint italicclass="text-teal">">//api.djinn.dev/v1/analyze",
    headers={class="text-teal">"X-API-Key": class="text-teal">"djinn_live_..."},
    json={
        class="text-teal">"candidate": {
            class="text-teal">"cv_file_url": class="text-teal">"https:class="text-text-faint italicclass="text-teal">">//storage.example.com/resume_martin.pdf",
            class="text-teal">"name": class="text-teal">"Sophie Martin",
        },
        class="text-teal">"job": {
            class="text-teal">"title": class="text-teal">"Senior Python Developer",
            class="text-teal">"skills": [class="text-teal">"Python", class="text-teal">"FastAPI", class="text-teal">"PostgreSQL", class="text-teal">"Docker"],
            class="text-teal">"experience_years": 5,
            class="text-teal">"location": class="text-teal">"Paris",
        },
    },
)

result = response.json()
print(fclass="text-teal">"Overall score: {result['score']}/100")
for dim in result[class="text-teal">"dimensions"]:
    print(fclass="text-teal">"  {dim['name']}: {dim['score']}/100 (weight: {dim['weight']}%)")
print(fclass="text-teal">"Recommendation: {result['recommendation']}")

Technical specifications

FormatsPDF, DOCX, TXT, JSON
Max size10 MB
P95 latency< 5s
SLA99.9%

Ready to integrate this service?

Test the API free with 1,000 calls included. Integrate in minutes.