🤖 Codweb Lab AI Mentor Agent

A production-grade, highly scalable, and intelligent AI Mentor platform built using FastAPI, Gemini Pro (Google GenAI), ChromaDB, and LangGraph. This project implements a strict modular architecture where the backend, frontend, asynchronous background workers, and the AI reasoning engine are completely decoupled.

---

🎯 What Does This AI Mentor Agent Do?

The Codweb Lab AI Mentor Agent is designed to act as a 24/7 autonomous technical guide for students and IT professionals. Instead of just answering generic questions, it uses advanced RAG (Retrieval-Augmented Generation) and multi-agent reasoning to provide deep, context-aware mentorship.

Key Capabilities:

• 🧠 Technical Upskilling & Mentorship: Provides structured guidance, code reviews, and conceptual clarity for complex domains like Python, Data Analytics, Machine Learning, and Web Development.
• 📚 Context-Aware Document Reading (RAG): Users can upload technical PDFs, research papers, or documentation. The agent ingests this data into ChromaDB and answers highly specific queries based only on the provided context.
• 🛠️ Autonomous Debugging: Utilizing LangGraph, the agent doesn't just guess answers. It can autonomously execute tools, search for the latest documentation, and self-correct its logic before giving a final solution.
• 📈 Career & Learning Path Mapping: Helps professionals navigating career breaks or transitioning into new tech roles by generating step-by-step, customized learning roadmaps.
• ⚡ Lightning Fast Responses: By leveraging Redis caching and background Celery workers, it delivers sub-second responses for repeated architectural queries without hitting the LLM again.

---

🚀 Core Technologies (Tech Stack)

• Backend: FastAPI, Python 3.11+
• Database: PostgreSQL (with Asyncpg driver), SQLAlchemy, Alembic
• AI & LLM: Google Gemini Pro (gemini-1.5-pro), Langchain
• Vector Store (RAG): ChromaDB
• Caching & Message Broker: Redis
• Background Jobs: Celery
• Agentic Workflows: LangGraph
• Frontend: Streamlit
• Infrastructure: Docker, Docker Compose, GitHub Actions (CI/CD)

---

🗺️ Project Development Roadmap (Phase-Wise Breakdown)

This project is meticulously structured into 9 logical phases to ensure smooth, error-free development and to avoid early over-engineering:

🛠️ Phase 1: Foundation & Environment

• Objective: Establish the repository baseline and virtual environment.
• Details: Initialize venv, install required packages via requirements.txt (FastAPI, Langchain, Celery, etc.), and configure a robust .gitignore to prevent sensitive or junk files from entering version control.

🔐 Phase 2: Backend Core, Database & JWT Auth

• Objective: Secure the backend foundation.
• Details: Establish a non-blocking asynchronous connection to PostgreSQL using asyncpg. Create imperative SQLAlchemy ORM models (Users, Chats), manage structural migrations with Alembic, and secure all endpoints using asymmetric JWT tokens.

🧪 Phase 3: Testing & CI/CD

• Objective: Maintain code quality and automate regression testing.
• Details: Write atomic API tests using pytest and configure GitHub Actions (ci.yml) to automatically trigger validation checks upon every code push.

🧠 Phase 4: AI Core & RAG System (Gemini Pro)

• Objective: Inject document intelligence and semantic retrieval capabilities.
• Details: Integrate Google Gemini Pro as the core LLM. Build a Retrieval-Augmented Generation (RAG) engine by chunking uploaded PDFs and storing dense vector embeddings in ChromaDB.

⚡ Phase 5: Async Workers & Caching

• Objective: Offload heavy CPU tasks and accelerate query responses.
• Details: Introduce Celery to run PDF text extraction and embedding generation asynchronously. Deploy Redis to serve as both the Celery message broker and a high-speed caching layer for repetitive queries.

🤖 Phase 6: Agentic AI Workflows

• Objective: Evolve the AI from a simple chatbot into an autonomous reasoning mentor.
• Details: Utilize LangGraph to build multi-step self-correcting memory loops, empowering the model to analyze context, decide on external tool execution pathways (e.g., web search), and synthesize comprehensive mentoring plans.

🐳 Phase 7: Full Dockerization

• Objective: Encapsulate the architecture for absolute environment parity.
• Details: Write a multi-stage Dockerfile and docker-compose.yml to orchestrate the FastAPI service, PostgreSQL database, Redis cache, and Celery workers into a unified, isolated network.

🎨 Phase 8: Frontend UI (Streamlit)

• Objective: Construct a highly responsive chat dashboard.
• Details: Build a dynamic reactive chat interface in pure Python using Streamlit, avoiding frontend boilerplate while seamlessly communicating with the FastAPI backend via HTTP requests.

🚀 Phase 9: Production Deployment

• Objective: Launch the platform globally.
• Details: Configure immutable cloud environments (e.g., AWS, Render), provision secure secret vaults for .env variables, and deploy the fully containerized production stack.

---

💻 Getting Started (Local Setup)

Follow these instructions to run the project locally on your machine:

1. Clone the repository:

git clone [https://github.com/your-username/ai-mentor-agent.git](https://github.com/your-username/ai-mentor-agent.git)

2. Navigate to the project directory:

cd ai-mentor-agent

---

---

⚖️ License & Commercial Use

This project is open-sourced under the GNU AGPLv3 License.

You are free to use, modify, and distribute this software, provided that any derivative works or network services running this code are also open-sourced under the same AGPLv3 license.

🏢 Commercial Dual-Licensing: If your organization wishes to use AI Mentor Agent in a closed-source, proprietary, or commercial environment without the AGPLv3 obligations, a commercial license must be acquired. Please contact Codweb Lab for enterprise licensing and white-label deployments.

---