Agentic AI Bootcamp

A collection of production-oriented AI engineering projects built throughout the Agentic AI Bootcamp.

The projects focus on building practical AI systems using RAG, agentic workflows, LLMs, tool calling, structured outputs, and automation.

---

Tasks

Task 01 — StudyMate RAG

A Retrieval-Augmented Generation system for studying and querying educational content.

Key Concepts:

• RAG pipeline
• Document processing
• Chunking and retrieval
• Vector search
• LLM integration
• Context-aware question answering

---

Task 02 — ResearchFlow

A multi-agent research workflow for planning, researching, reviewing, and generating structured research reports.

Key Concepts:

• PydanticAI
• Multi-agent architecture
• Multi-model architecture
• Google Gemini
• OpenAI
• Tavily web search
• Structured outputs with Pydantic
• Research planning
• Research execution
• Critic / review loop
• Retry and quality control

---

Final Project — Campus Regulations Advisor

An agentic RAG system designed to answer questions about university policies and regulations using grounded information from official policy documents.

The system combines a Streamlit frontend, FastAPI backend, LangGraph agent workflow, LlamaIndex hybrid retrieval, and n8n automation.

Architecture

Streamlit Frontend
        │
        ▼
    FastAPI API
        │
        ▼
   LangGraph Agent
        │
   ┌────┴────┐
   ▼         ▼
Router    Retriever
             │
        ┌────┴────┐
        ▼         ▼
      BM25     Dense Search
        │         │
        └────┬────┘
             ▼
       Knowledge Base
             │
             ▼
         Generator
             │
             ▼
    Grounding Evaluator
             │
       ┌─────┴─────┐
       ▼           ▼
   High Score   Low Score
       │           │
       ▼           ▼
    Answer       HITL Review