🌄 ApplyVista — AI-Powered Job Application Assistant (MVP)

ApplyVista is an intelligent, offline-first job application assistant that generates ATS-ready skills, summaries, cover letters, recruiter outreach emails, follow-up emails, and short-form responses, all grounded in your own resume and project documents using a custom Retrieval-Augmented Generation (RAG) pipeline.

Built using FastAPI, Streamlit, LangChain, Ollama, and ChromaDB, ApplyVista runs completely locally, ensuring privacy, speed, and full control of your data.

---

🚀 Features

🔍 Retrieval-Augmented Generation (RAG)

• Processes and indexes your personal documents
• Creates embeddings using SentenceTransformers
• Stores vectors in ChromaDB for semantic search
• Retrieves the most relevant snippets for grounded LLM outputs
• Avoids hallucinations by anchoring responses in your real content

🧠 Local LLM (Fully Offline)

• Powered by Ollama Llama 3.2 (3B)
• No API keys
• No cloud calls
• No cost
• 100% private

🧩 Job Application Generation

Automatically produces:

• Key Skills & Keywords
• ATS-friendly Resume Summary
• Full Cover Letter
• Application Email
• Recruiter Outreach Email
• Follow-up Email
• “Why this Job?” Short Response
• Skill Alignment (Have vs. Gap)

💬 Chat With Your Documents

Ask personalized questions such as:

• “Summarize my machine learning project.”
• “Rewrite my experience for a senior AI engineer role.”
• “Give me resume bullet points for my RAG pipeline.”

🖥️ Streamlit Interface

• Paste job description
• Click Generate Application Content
• Edit each section in its own tab
• Download results as DOCX or PDF

⚙️ FastAPI Backend

• /jd → Full job application package
• /chat → Q&A grounded in your documents
• Auto-generated Swagger docs at: http://localhost:8000/docs

---

🧰 Tech Stack

Component	Purpose
FastAPI	Backend REST API
Streamlit	Frontend UI
Ollama	Local LLM runtime
LangChain	RAG orchestration
ChromaDB	Vector storage + retrieval
FPDF/DOCX	Export utilities
Docker	Optional containerization

---

⚙️ Setup Instructions

1️⃣ Clone the Repository

git clone https://github.com/sheikhmunim/ApplyVista.git
cd ApplyVista

1️⃣ Clone the Repository

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

3️⃣ Install Dependencies

pip install -r requirements.txt

4️⃣ Install & Start Ollama

ollama pull llama3.2:3b ollama serve

▶️ Run the Application

uvicorn app.api:app --reload

API Docs: http://localhost:8000/docs

Start Streamlit Frontend

streamlit run app/app_fast.py

User Interface: http://localhost:8501

🧠 Example Workflow

Step 1 — Paste the Job Description

Insert any job description into the input box.

Step 2 — Click Generate Application Content

The system will automatically process the description using the NLP pipeline.

Step 3 — Receive Automatically Generated Outputs

You will instantly get:

• Skills & Keywords
• ATS Summary
• Cover Letter
• Application Email
• Recruiter Outreach Email
• Follow-up Email
• Short “Why this Role?” Answer
• Skill Alignment Report

Step 4 — Edit Text Directly in Streamlit

All generated content is displayed with editable text fields for immediate adjustments.

Step 5 — Download as DOCX or PDF

Export your final application package in your preferred format.

---

🛠️ Implementation Notes

📌 RAG Pipeline

• Loads and chunks resume and project documents
• Converts text into vector embeddings
• Performs top-K similarity search to find the most relevant context
• Builds structured contextual prompt blocks
• Generates each component (cover letter, emails, summaries, etc.) using carefully designed prompts

📌 FastAPI Backend

• Clean, well-typed endpoints using Pydantic models
• Stateless design suitable for Docker and cloud deployment
• Separate handlers for:
  • Job description (JD) processing
  • General chat / utility endpoints

📌 Streamlit Frontend

• Multi-tab, in-browser editor for all generated content
• Real-time updates using Streamlit session state
• Export functionality to download results as PDF or DOCX

🐳 Docker Support (Optional)

Build Docker Image

docker build -t applyvista .

Run Docker Container

docker run -p 8000:8000 -p 8501:8501 applyvista

🗺️ Future Enhancements

• Resume–JD Matching Score
• Personalized Job Recommendations
• Improved reranking with BGE/ColBERT
• Resume optimization engine
• Skill gap analysis visuals
• Multi-user account system
• Analytics dashboard

🪪 License This project is licensed under the MIT License.

✨ Author Sheikh Abdul Munim
Master of Artificial Intelligence — RMIT University

🔗 LinkedIn: https://www.linkedin.com/in/sheikh-abdul-munim-b19391158
🔗 GitHub: https://github.com/sheikhmunim