← All courses
BSAI-101 · Beginner

Agents & LLMs

From 'what is a token?' to shipping your own AI agent — in two weekends.

A 4-day live course that gives non-coder professionals a solid foundation in LLMs and AI agents. You'll understand how these systems work, learn to talk to them effectively, and build a working agent for your own domain using AI coding tools.

Join the WaitlistSee the Syllabus
Duration
4 days · 2 weekends · ~11 hours
Level
Beginner
Format
Live on Zoom · 2 consecutive weekends with a gap week
Who is this for

Built for professionals, not programmers.

Working professionals who want to understand and build with AI — no coding experience required.

What you'll walk away with

By the end of the course.

✓

A solid mental model of how LLMs and AI agents work

✓

A working capstone agent built for your own domain

✓

The vocabulary to evaluate AI products and build your own

✓

Hands-on experience with AI coding tools (Cursor / Claude Code)

Syllabus

Four days, one clear arc.

Each day builds on the last. By Day 4 you ship a working agent.

01

What is an LLM, really?

Build your mental model of how LLMs work — tokens, prediction, context, temperature.

  • ●Next-token prediction & tokenisation
  • ●Context windows & temperature
  • ●Why LLMs hallucinate
  • ●The model landscape: closed vs open
  • ●Live demos: tokenizer, model compare, cost calculator
02

Talking to LLMs well

Learn to steer LLMs with prompting techniques and context engineering.

  • ●System vs user prompts
  • ●Few-shot & role prompting
  • ●Structured output
  • ●Context engineering: right info, right shape, right time
  • ●Chain of thought reasoning
03

From LLM to Agent

Turn an LLM into an agent with tools, a loop, and automated testing.

  • ●What makes an LLM an agent
  • ●Tool calling: define, decide, execute
  • ●The agent loop
  • ●Building a Streamlit chat agent
  • ●Evals: testing AI output
04

Ship your agent

Add memory, connect to external tools via MCP, and ship your capstone.

  • ●Short-term & long-term memory
  • ●MCP: the universal tool protocol
  • ●Capstone work session with Cursor / Claude Code
  • ●Agent safety & regulation
  • ●Where to go next
Capstone Project

Your Domain Agent

A working Streamlit chat agent extended for your domain — with tools, memory, and evals. Built using AI coding tools, not manual coding.

Prerequisites

What you need before you join.

  • ●No coding experience required
  • ●A laptop with internet access
  • ●Willingness to experiment and build
Tools we'll use

Bring these, or install during setup.

Anthropic APIPython (guided setup)Cursor or Claude CodeStreamlit
Reserve your spot

Join the waitlist for BSAI-101: Agents & LLMs.

We'll email you when the next cohort opens. Early signups get launch pricing.