🤖 Job Application Assistant

An AI-powered multi-agent system that generates a complete job application package — deployed as a cloud-native application with Docker, OpenShift, Ansible, and GitHub Actions CI/CD.

🌐 Live Demo

🎯 What it does

Input a company name, job role, and your name — 4 AI agents collaborate to generate:

Output	Description
🔍 Company Research Report	Culture, values, requirements, recent news
📄 Personalized Cover Letter	250–350 words, tailored to the company
🎯 Interview Prep	5 likely questions with tips on how to answer
📧 Formal Cold Email	Ready to send to the hiring manager on LinkedIn

---

🛠️ Tech Stack

AI & Application

Tool	Purpose
CrewAI	Multi-agent orchestration framework
Groq (LLaMA 3.3 70B)	LLM for agent reasoning and writing
Serper API	Real-time web search tool
LiteLLM	LLM provider interface
Streamlit	Interactive web UI
Python 3.11	Core language

DevOps & Cloud

Tool	Purpose
Docker	Containerization
Docker Hub	Image registry
OpenShift	Enterprise Kubernetes platform
Ansible	Deployment automation
GitHub Actions	CI/CD pipeline

---

🏗️ Project Structure

job-research-agent/
├── src/
│   ├── __init__.py
│   ├── agents.py           # 4 Agent definitions
│   └── tasks.py            # 4 Task definitions
├── k8s/
│   ├── deployment.yaml     # OpenShift deployment manifest
│   ├── service.yaml        # Service manifest
│   └── route.yaml          # Route manifest for public URL
├── ansible/
│   └── playbook.yml        # Ansible deployment playbook
├── .github/
│   └── workflows/
│       └── deploy.yml      # GitHub Actions CI/CD pipeline
├── app.py                  # Streamlit UI
├── Dockerfile              # Container definition
├── requirements.txt        # Python dependencies
├── .env                    # API keys (not pushed to GitHub)
├── .gitignore
└── README.md

---

🚀 How it works

AI Pipeline

User Input (Company + Role + Name)
            ↓
Researcher Agent → Searches web → Research Report
            ↓
Writer Agent → Reads research → Cover Letter
            ↓
Coach Agent → Reads research → 5 Interview Questions + Tips
            ↓
Emailer Agent → Reads research → Formal Cold Email
            ↓
All 4 outputs displayed in Streamlit tabs

DevOps Pipeline

git push → GitHub Actions triggers
            ↓
Docker build → Docker Hub push
            ↓
oc apply manifests → OpenShift deployment
            ↓
Live on OpenShift public URL

---

⚙️ Setup & Installation

Local Development

1. Clone the repository

git clone https://github.com/YashRM27/job-research-agent.git
cd job-research-agent

2. Install dependencies

pip install -r requirements.txt

3. Create a .env file

GROQ_API_KEY=your_groq_api_key_here
SERPER_API_KEY=your_serper_api_key_here

4. Run the Streamlit app

streamlit run app.py

---

Docker

docker build -t job-research-agent .
docker run -p 8501:8501 --env-file .env job-research-agent

---

OpenShift Deployment

# Login to OpenShift
oc login --token=YOUR_TOKEN --server=YOUR_SERVER

# Create secrets
oc create secret generic api-keys \
  --from-literal=GROQ_API_KEY=your_key \
  --from-literal=SERPER_API_KEY=your_key

# Apply manifests
oc apply -f k8s/deployment.yaml
oc apply -f k8s/service.yaml
oc apply -f k8s/route.yaml

# Check status
oc get pods
oc get route

---

Ansible Deployment

ansible-playbook ansible/playbook.yml

---

🔄 CI/CD Pipeline

Every git push to main automatically:

1. Builds Docker image
2. Pushes to Docker Hub
3. Deploys to OpenShift
4. Verifies deployment status

Required GitHub Secrets

DOCKERHUB_USERNAME
DOCKERHUB_TOKEN
OPENSHIFT_TOKEN
OPENSHIFT_SERVER

---

🔑 Get Free API Keys

• Groq API (Free LLM): https://console.groq.com
• Serper API (Free web search): https://serper.dev

---

🔮 Future Improvements

• RAG — upload resume PDF for personalized output
• Support multiple output formats (PDF, DOCX)
• Email sending integration
• Job fit score based on resume vs job description
• Fix CrewAI non-root permission conflict on OpenShift
• Add persistent volume for ChromaDB storage
• Horizontal pod autoscaling
• Multi-environment deployment (dev / staging / prod)