💼 Job Application Assistant

An AI-powered tool that analyses your resume against any job description and generates a complete application package — in under 30 seconds.

What it does

Upload your resume PDF and paste a job description. The pipeline runs four AI chains and returns:

• Skill Gap Report — match score, missing skills with learning resources, your strengths for the role
• 3 Cover Letter Versions — formal, conversational, and concise — each tailored to your actual experience
• Interview Preparation — 10 role-specific questions with suggested answers based on your resume
• Downloadable ZIP — all outputs bundled for offline use

Demo

Upload resume → paste JD → click Analyse → download your package

Tech stack

Layer	Technology
LLM	Llama 3.3 70B via Groq API
Orchestration	LangChain LCEL
Structured output	Pydantic v2
PDF parsing	PyMuPDF
UI	Streamlit
Language	Python 3.10+

How the pipeline works

Resume PDF + Job Description
        ↓
Step 1 — Parallel (simultaneous):
   ├── JD Analysis      → extracts skills, requirements, culture signals
   └── Resume Analysis  → extracts skills, projects, achievements

Step 2 — Skill Gap Analysis
   → compares JD requirements vs resume capabilities
   → generates match score + prioritised gap list

Step 3 — Cover Letter Generation
   → 3 versions using candidate's actual experience

Step 4 — Interview Preparation
   → 10 questions with personalised suggested answers

Steps 1 runs JD and resume analysis simultaneously using RunnableParallel — cutting processing time by ~50%.

Project structure

job-application-assistant/
├── app.py              # Streamlit UI
├── core/
│   ├── models.py       # Pydantic output models
│   ├── parser.py       # PDF and text parsing
│   └── chains.py       # LangChain LCEL pipeline
├── utils/
│   └── helpers.py      # Formatting and file saving
└── outputs/            # Generated files (gitignored)

Run locally

1. Clone the repo

git clone https://github.com/thammi21/resumatch-ai
cd job-application-assistant

2. Create virtual environment

python -m venv venv
venv\Scripts\activate        # Windows
source venv/bin/activate     # Mac/Linux

3. Install dependencies

pip install -r requirements.txt

4. Add your API key

Get a free Groq API key at console.groq.com

Create a .env file:

GROQ_API_KEY=your_key_here

5. Run

streamlit run app.py

Key concepts demonstrated

• LCEL pipe syntax — prompt | llm | parser chains
• RunnableParallel — JD and resume analysis run simultaneously
• PydanticOutputParser — structured validated output at every step
• RunnableLambda — Python functions as Runnable chain components
• Context window management — automatic truncation for long documents
• Dual LLM instances — temperature=0 for analysis, temperature=0.7 for creative writing

What I learned building this

• How to design a multi-step LangChain pipeline where each step's output feeds the next

• Why parallel chains matter — analysis time dropped from ~12s to ~6s

• How PydanticOutputParser injects format instructions directly into prompts

• The difference between RunnableLambda (any Python function) and RunnablePassthrough (preserve inputs while adding new keys)

• How to use callbacks for real-time progress tracking in Streamlit

  🔴 Live Demo → resumatch-ai.streamlit.app

⚠️ App may take 30–60 seconds to wake up if not visited recently.