📍 In-Person Workshop • GIKI Campus
atomcamp, in collaboration with GIK, is organizing a 2-day in-person workshop
Click here to confirm the registration fee: https://giki.edu.pk/agentic-payment/
atomcamp × Ghulam Ishaq Khan Institute (GIKI) — A 2-day hands-on program
🏠 Accommodation available:
- GIKI Hostel (PKR 2,000/room/night)
- or choose the GIKI Guest House (PKR 5,500/room/night).
This hands-on workshop focuses on Agentic AI architectures and Retrieval-Augmented Generation (RAG). It is designed for students, developers, and data professionals who want to move beyond basic LLM API usage and learn how to build context-aware AI systems using LangChain and modern agent frameworks.
Across two days, you’ll learn how modern AI agents think, plan, remember information, and use tools. You’ll also build RAG pipelines that allow AI models to answer questions using your own data instead of guessing. By the end of Day 2, you will have built a working Knowledge Agent that can understand and respond to queries from a private dataset.
⚠️ Prerequisite: Proficiency in Python and basic understanding of LLM APIs (OpenAI / Anthropic).
Here’s What You’ll Learn
✓ How modern AI agents reason, plan, and act
✓ Designing memory-aware AI systems
✓ Building RAG pipelines for private data
✓ LangChain & LCEL best practices
✓ Architecting AI systems that scale beyond demos
Curriculum
Click each module to expand the full breakdown.
The Shift from RAG to Autonomy: Why Agents?
- From static chatbots to autonomous AI systems
- Motivation behind Agentic AI and enterprise value
- Limitations of traditional RAG pipelines
The P-P-A-R-M Framework
- Perception: Understanding inputs and environment
- Planning: Task decomposition and goal setting
- Action: Tool usage and execution
- Reflection: Self-evaluation and correction
- Memory: Short-term and long-term context
Architectural Foundations
- Layers of reasoning, orchestration, and execution
- Designing scalable agent workflows
- Separation of concerns in agent systems
Hands-On Lab: Build and configure a basic AI agent with defined identity, goals, and structured prompts.
LangChain Core Concepts
- LangChain Expression Language (LCEL)
- Building composable AI chains
- Prompt templates and orchestration
- Managing state with ChatMessageHistory
RAG Data Pipeline
- Document loading and preprocessing
- Text chunking strategies
- Vector embeddings and similarity search
- Context injection into prompts
Vector Databases
- Introduction to vector databases
- Implementing ChromaDB
- Indexing and retrieval workflows
Hands-On Lab: Create a Private Knowledge Bot that answers questions from uploaded documents using semantic search.
Tool Binding & Agent Types
- Converting Python functions into AI tools
- Pydantic validation for tool inputs
- Zero-shot, Conversational, and Structured agents
Error Handling & Self-Correction
- Tool invocation strategies
- Self-correction loops
- Handling failed actions gracefully
Hands-On Lab: Develop an AI agent that invokes external APIs and executes real-world tasks with validation.
Agent Roles & Collaboration
- Defining agent roles and backstories
- Researcher, Writer, Reviewer personas
- Context sharing between agents
Task Orchestration
- Sequential and hierarchical workflows
- Designing collaborative pipelines
- Coordinating multiple agents toward one goal
Hands-On Lab: Build a multi-agent content engine for research, drafting, and automated review.
Containerization & Deployment
- Dockerizing the backend (FastAPI) and frontend (Streamlit / web UI)
- Environment configs: .env, secrets, and build-time vs run-time variables
- Local-to-prod parity (same setup across environments)
Cloud Deployment Strategies
- Serverless backend deployment (functions / APIs)
- Static hosting for frontend (fast delivery + scalable)
- Managing API keys, rate limits, and secure access in production
Monitoring & Observability
- Logging best practices for agent workflows
- Request tracing across the stack (agent → tools → APIs)
- Basic monitoring for latency, failures, and cost tracking
Final Project: Deploy the full-stack Agent Web App to a production environment — live endpoint + working UI + monitoring enabled.
Who This Is For
- Students & Fresh Graduates: Building AI-powered products and gaining hands-on experience
- Professionals: Wanting to upskill fast in agentic AI and modern architectures
- Career Changers: Seeking career growth through practical AI tools
- Freelancers & Entrepreneurs: Building AI-driven products, services, and side-projects
Meet Our Incredible Trainers
Join Now & Build Your Own Agent
Limited seats available. Secure your spot for the 2-day in-person experience at GIKI Campus.
