🔥 October 24 & 25, 2026 • Live Online

Build, Secure & Ship Your Own AI Agent in 2 Days

Hands-on AI Bootcamp with Python, in Tamil. Learn to build intelligent agents, give them memory and tools, and connect them to the real world.

📅 Oct 24 & 25 (Sat/Sun) ⏰ 10:00 AM to 3:00 PM IST, with lunch and breaks. 🎙️ Tamil 💵 ₹500 Only 🔴 Recordings Included
Register Now Explore Syllabus

Bootcamp starts in:

00
Days
00
Hours
00
Mins
00
Secs
> python run_agent.py
[SYSTEM] Initializing Hermes Agent...
[AGENT] Tools loaded: WebSearch, Calculator, ScriptRunner
> Task: Analyze latest tech news, format as JSON, and send to Telegram.
[AGENT] Thinking...
🎙️ Tamil Sessions
📼 Recordings Included
💻 60% Hands-on Labs
🛠️ Free Tools Only
🌱 Beginner-friendly

What You'll Build

Stop talking about AI. Start building it.

🧠

Custom AI Agent

Build an agent in Python step-by-step. Implement the Think-Act-Observe loop, give it memory, and connect it to custom tools.

🤖

Hermes Agent

Deploy a powerful Hermes Agent (by Nous Research) that runs persistently, uses skills, schedules tasks, and lives on Telegram.

🛡️

Guarded Support Bot

Build a secure bot that blocks prompt injections, handles suspicious messages, and automatically routes tasks using Laya guardrails.

Detailed Syllabus

Format: Online, in Tamil, with recordings.
Prerequisite: Basic programming knowledge (any language).

DAY 1 (Oct 24): Python, LLMs & Agent Fundamentals

🎯 Goal: Understand how an AI agent works internally, then meet a production-grade one.

Session 1: Python for AI

What you'll learn:
  • Setting up Python, virtual environments, and pip.
  • Functions, lists, and dictionaries (agent building blocks).
  • Reading and writing JSON.
  • Calling web APIs using the requests library.
  • Storing API keys safely in a .env file.
💻 Lab: Write a script that calls a public API, parses the JSON response, and prints a clean summary.
✅ Outcome: A working Python environment and the API skills needed for the bootcamp.

Session 2: How LLMs Work

What you'll learn:
  • What an LLM is, in plain Tamil, without heavy math.
  • Tokens: how text is split, and cost limits.
  • The context window: what the model "sees" at once.
  • Temperature and sampling.
  • Hallucinations and choosing the right model (free vs paid).
💻 Lab: Run the same prompt at different temperatures and compare the results.
✅ Outcome: Clear mental model of what an LLM can and cannot do.

Session 3: Calling APIs & Chatbots

What you'll learn:
  • Making your first LLM API call from Python.
  • Message roles: system, user, and assistant.
  • Writing effective system prompts.
  • Keeping conversation history for memory.
  • Handling API errors and rate limits gracefully.
💻 Lab: Build a command-line chatbot with a custom personality.
✅ Outcome: Your first working AI application.

Session 4: Structured Output & Tools

What you'll learn:
  • Prompt engineering: instructions, examples, constraints.
  • JSON mode and enforcing structured output.
  • Why structured output matters for code.
  • Tool (function) calling: letting the LLM decide when to run code.
  • Defining tool descriptions and parameters.
💻 Lab: Build a script where the LLM calls a calculator and weather tool.
✅ Outcome: The core mechanism that turns a chatbot into an agent.

Session 5: Build an Agent from Scratch

What you'll learn:
  • The agent loop: think → act → observe → repeat.
  • Wiring multiple tools into one agent.
  • Short-term vs summary memory.
  • Stopping conditions to prevent infinite loops.
  • Error handling when a tool fails.
💻 Lab: Build your own mini agent with at least two tools and memory.
✅ Outcome: A hand-built agent, understanding what frameworks do under the hood.

Session 6: Meet Hermes Agent

What you'll learn:
  • What Hermes Agent is and how it differs from your build.
  • Installation and setup via hermes setup.
  • Connecting a model provider (free-tier or local).
  • Running your first automated tasks.
💻 Lab: Give the same task to your agent and Hermes; compare results.
✅ Outcome: A running Hermes installation ready for Day 2.
📝 Day 1 Homework: Download the Laya model files and confirm Hermes runs perfectly on your machine.

DAY 2 (Oct 25): Hermes, Laya Guardrails & Capstone

🎯 Goal: Make the agent remember, automate, stay safe, and serve real users.

Session 7: Hermes Deep Dive

What you'll learn:
  • Persistent memory across multiple sessions.
  • Skills: reusable abilities the agent can learn.
  • Creating your own custom skill.
  • Scheduled tasks (e.g., automated daily summaries).
💻 Lab: Write a custom skill and set up a scheduled daily task.
✅ Outcome: An agent that remembers and works without being asked.

Session 8: Telegram Gateway

What you'll learn:
  • Why meeting users on messaging apps matters.
  • Creating a Telegram bot and getting tokens.
  • Connecting Hermes to Telegram.
  • Basic access control (who can talk to your agent).
💻 Lab: Chat with your own AI agent directly from your phone.
✅ Outcome: A live, accessible Telegram AI assistant.

Session 9: System 1 vs 2 AI Thinking

What you'll learn:
  • System 2 (slow LLMs) vs System 1 (fast classifiers).
  • Why sending everything to an LLM is expensive and unsafe.
  • Introduction to Laya: fast decision models returning typed answers.
  • Where Laya fits: gatekeeper, not the agent's brain.
💻 Lab: Classify messages using an LLM vs Laya; compare speed.
✅ Outcome: Know when to use a fast classifier vs an LLM.

Session 10: Laya Guardrails & Routing

What you'll learn:
  • Prompt-injection and jailbreak detection.
  • Message triage (urgent, billing, spam).
  • Model routing: cheap vs stronger models based on difficulty.
  • Confidence thresholds and logging decisions.
💻 Lab: Add a guard step and a router in front of your agent.
✅ Outcome: A safer, faster, and cheaper agent pipeline.

Session 11: Capstone Project

What you'll build: A Telegram Support Bot that:
  • Blocks suspicious/injection-style messages via Laya.
  • Triages incoming messages by category & urgency.
  • Routes simple queries to cheap models, complex to strong models.
  • Remembers user history and sends daily summaries.
✅ Outcome: A portfolio-ready project you can demo and put on GitHub.

Session 12: Wrap-up & Next Steps

What you'll learn:
  • Security best practices (sandboxing, API keys).
  • Cost control and rate-limit handling.
  • Demo time: show off your capstone project!
  • Roadmap: RAG, multi-agent setups, deployment.
✅ Outcome: A clear path for what to learn and build next in your AI journey.

🎁 What You Take Home

  • ✅ A hand-built Python agent with tools and memory
  • ✅ A Hermes-powered Telegram assistant
  • ✅ A Laya-guarded support bot with triage and routing
  • ✅ Starter code, lifetime session recordings
  • ✅ A roadmap for what to learn next

⚡ Skills Gained

Python for AI LLM APIs Prompt Engineering Structured Output Tool Calling Agent Design Agent Memory Messaging Integration Guardrails Model Routing Agent Security Basics

👨‍💻 Who is this for?

Students, developers, and IT professionals who know basic programming (loops, functions, variables) and want to transition from basic prompt engineering to building robust, autonomous AI agents.

⚙️ Tools Needed

  • • A laptop with Python 3.10+ installed
  • • A code editor (VS Code recommended)
  • • A Telegram account for bot testing
  • • A free API key (Groq or OpenRouter)
  • • Laya model files (download link provided)

Limited Seats Available

Bootcamp Pass

₹500 / both days
Register Now

Frequently Asked Questions

No prior AI experience is needed. However, basic programming knowledge (like writing functions, loops, and variables in Python or any other language) is highly recommended to keep up with the labs.

Yes, all registrants will receive lifetime access to the full recordings after the bootcamp concludes.

The entire bootcamp will be taught in Tamil to ensure maximum clarity and deep understanding.

No! We will strictly use tools and APIs that offer generous free tiers (like Groq, OpenRouter, and local models) so you won't need to spend extra money during the bootcamp.

Don't worry, you can always catch up using the session recordings provided to all registered participants.

Ready to build your first AI Agent?

Seats are strictly limited to ensure interactive Q&A. Reserve your spot today.

Claim Your Spot - ₹500