Resume Intelligence Engine

01 —

System Architecture · 3D View

02 —

Architecture Diagram

JD Input

JSON / Text

↓

FastAPI Router

REST Endpoints

↓

JD Parser

NLP · Regex

Suitability

Scoring · Ranking

Skill Bridge

Prioritization

Profile Store

Candidate Data

↓

Generator

Content Builder

↓

PDF Export

ReportLab

DOCX Export

python-docx

JSON Response

REST API

03 —

Screenshots & Output

terminal

$ uvicorn main:app --reload --port 8000

INFO: Uvicorn running on http://127.0.0.1:8000

POST /api/generate-resume HTTP/1.1

✓ JD parsed: 14 skills extracted

✓ Suitability score: 87.3%

✓ Skills prioritized: [FastAPI, NLP, Python...]

✓ Summary generated: 68 words

✓ PDF rendered → resume_darshan_joshi.pdf

✓ DOCX rendered → resume_darshan_joshi.docx

Response time: 1.43s Status: 200 OK

$

API Terminal Output

FastAPI server live generation log

Skill Match Scores

Python98%

FastAPI94%

NLP Match87%

PDF Rendering82%

Regex Parser90%

Skill Match Scores

Per-skill suitability scoring

Data Output

{
"suitability_score": 0.873,
"matched_skills": ["FastAPI", "NLP", "Python"],
"missing_skills": ["Docker"],
"outputs": {
"pdf": "/out/resume.pdf",
"docx": "/out/resume.docx"
}
}

JSON API Response

Structured response payload

Project Structure

📁 resume-intelligence-engine/

├─ main.py FastAPI entry

├─ jd_parser.py NLP pipeline

├─ suitability.py Scoring engine

├─ skill_bridge.py Prioritization

├─ generator.py Content builder

├─ exporters/

│ ├─ pdf_export.py ReportLab

│ └─ docx_export.py python-docx

└─ requirements.txt

Project Structure

Modular file organization

04 —

What I Built

›

Architected and implemented a modular FastAPI backend for dynamic resume and cover letter generation from JDs.

›

Designed an NLP-driven JD parsing pipeline using regex and heuristic classification to extract skills, responsibilities, and role families.

›

Engineered dynamic skills prioritization and experience reordering mechanism for recruiter relevance.

›

Developed rule-based intelligence for summary auto-generation and project relevance scoring.

›

Built scalable PDF/DOCX generation workflow using ReportLab and python-docx.

›

Structured codebase into independent modules: profile store, suitability engine, skill bridging, generator.

05 —

Project Insights

✎ Personal Notes & Learnings

Markdown Editor

Live Preview

What I Built

A modular FastAPI backend that generates tailored resumes on-the-fly by parsing job descriptions with NLP.

Key Challenges

Designing a robust jd_parser that handles unstructured text reliably
Getting the skill prioritization order right without manual rules
Making PDF layout consistent across different content lengths

What I Learned

Regex + heuristics can outperform heavy NLP models for well-scoped tasks
Modular architecture pays off immediately when adding new export formats

Next Steps

Add an LLM-powered summary rewriter as an optional layer
Build a frontend UI for non-technical users

✓ Insights saved locally