Llms In Practice For Students

LLMs in Practice — From Prompt to Pipeline

Hands-on course materials for computer engineering students. This course takes you from your first API call to building a fully autonomous NER pipeline using an AI agent — in two focused modules.

---

What you will learn

**Module 1 — Prompt Engineering** Understand how LLMs work and learn the techniques that make the difference between a prompt that kind of works and one that reliably produces the output you need.

**Module 2 — Agentic AI Coding** Move beyond the chatbox. Learn how agents work, how tools give models access to the real world, and how to use Claude Code to build a complete NLP project from scratch using only natural language.

---

Course structure

Module 1 — Prompt Engineering
├── 00_introduction.ipynb       — setup, API key, course overview
├── 01_getting_started.ipynb    — first API call, roles, get_completion helper
├── 02_prompting_tips.ipynb     — 5 essential tips every NLP engineer needs
└── 03_prompting_examples.ipynb — zero-shot, few-shot, chain-of-thought

Module 2 — Agentic AI Coding
└── Covered in the live classroom session using Claude Code CLI
    - What is an agent
    - The agent loop: perceive → plan → act → observe
    - Tool use and how models interact with the real world
    - CLAUDE.md: giving the agent memory
    - Live demo: building an NER pipeline from scratch

---

Prerequisites

• Basic Python (functions, loops, dictionaries, working with JSON)
• No prior NLP or AI experience required
• No machine learning background needed

---

Stack

Tool	Purpose
Python 3.11	Language used throughout
OpenRouter	Unified API gateway — one key for all models
openai SDK	Client library (OpenRouter is OpenAI-compatible)
python-dotenv	Loading environment variables
pytest	Testing in Module 2

**Why OpenRouter?** One API key gives access to Claude, GPT-4, Llama, Mistral and more. You can switch mode