Resume Readiness Intelligence Engine

AI assistant that analyzes a résumé and job description, detects skill gaps, retrieves learning resources, and generates a 4-week study plan.

Quickstart

Requirements: Python 3.11+ and a Groq API key.

1. Install deps

   python -m pip install --upgrade pip
   python -m pip install -r requirements.txt

   (Or make install.)

2. Configure env (.env in repo root)

   GROQ_API_KEY=your_key_here
   # Optional overrides: RAG_MODEL, RAG_PERSIST_DIR, RAG_COLLECTION_NAME, RAG_EMBEDDING_MODEL, RAG_TOP_K, RESUME_PARSER_MODEL, RESUME_PARSER_TEMPERATURE
   

3. Run backend

   uvicorn backend.app.api.main:app --host 0.0.0.0 --port 8000 --reload

   (Or make run-backend.)

4. Serve frontend

   python -m http.server 3000 --directory frontend

   Visit http://localhost:3000 and upload a résumé PDF + paste a job description.

How it works (pipeline)

1. Résumé → XML (resume_parser.py)
2. Job skills (LLM) → <jobSkills> XML (job_skill_eval.py)
3. Résumé skills (LLM) → <skillsEvaluation> XML (resume_skill_eval.py)
4. Skill gaps → <skillGaps> XML (skill_gap_eval.py)
5. RAG study plan: retrieve from Chroma (VectorDB/), generate 4-week plan (rag_agent.py)
6. API (backend/app/api/main.py) returns:
   • study_plan (markdown), plan_structured (weeks/tasks/resources), plan_reasoning
   • job/resume skills and gap XML

API endpoints

• GET /health — uptime probe
• POST /analyze — accepts JSON { resume_pdf_path, job_description }
• POST /analyze-upload — multipart form (resume_pdf, job_description) used by the frontend

Frontend

Static HTML/CSS/JS (frontend/index.html), no build step. Fetches POST /analyze-upload. Views: Study Plan, Structured, Raw, Reasoning.

Project structure

backend/app/api/              # FastAPI entrypoint
backend/app/workflow/         # Orchestrator wiring services
backend/app/services/         # Resume/job skill eval, skill gap, RAG agent
backend/app/utils/            # XML helpers
frontend/                     # Static UI
VectorDB/                     # Chroma persistence (local)
docs/                         # Brief, appendix, deliverables checklist
demo/                         # Examples of resumes and job descriptions for local tests and demo
Makefile                      # install/run helpers
requirements.txt

Deliverables

======= Planned expansion:

• Software Engineering

• Finance

• Mechanical Engineering

Additional technical and non-technical career paths

• Both entry-level and senior candidates

MVP Features

1. Skill Gap Detection

Users upload their resume or answer domain-specific questions.

• Our system:

  • Analyzes skills, tools, and domain knowledge

  • Compares them to real job requirements

  • Returns structured “knowledge gap” outputs

2. RAG-Based Knowledge Retrieval

The RAG database is built from credible sources:

• ML & Statistics textbooks

• Ace the Data Science Interview

• Applied DS/ML/AI resources

• Articles, videos, PDFs, and tutorials (Domain Specific)

The engine retrieves explanations, examples, and runnable insights—not generic fluff.

3. Personalized Study Plans

For every knowledge gap, the system generates:

• Sequenced study modules from videos, articles, and textbook excerpts

• Timelines and difficulty progression

High-Level Architecture

Frontend: Built using HTML and JavaScript. Communicates with backend through Fast API calls.

Backend:

• Fast API, Python

• Inference - Groq

• Model - llama-3.1-8b-instant

RAG Pipeline:

• Textbook into markdown using Cloud GPUs

• Chunking & preprocessing by header

• Sentence transformer embedding (all-MiniLM-L6-v2)

• Storage in vector DB (ChromaDB)

• Langchain

• Retrieval (Precision(k=3)) + Generation

Agentic Workflow

Agent 1 [PDF converter]
Input: Resume
Output: XML of resume

Goes to:


Agent 2[Job description Skills]
Input: Job description
Output: Skill and proficiency requirements for the specific job

Goes to:

Agent 3[Resume Skills]
Input: XML of resume
Output: Skill and proficiency requirments with respect to the skills outlined from agent 2


Goes to:


Python function to get the gaps
Input: Skills 
Output: Gaps

Goes to:


Agent 4[RAG + Study Plan]
Input: Skill gaps
Output: Study plan to get proficient in the necessary skills

• Working prototype (steps above; Makefile included)
• Product brief: docs/product_brief.md
• Technical appendix: docs/technical_appendix.md
• Deliverables checklist: docs/deliverables.md
• Add your deck and demo video to docs/ or link them in docs/deliverables.md

Notes

• Defaults use HuggingFace all-MiniLM-L6-v2 embeddings with Chroma. Override paths/models via env vars if needed.
• main is protected; merge changes via PR.