From curious
to expert.

What AI is, how it works, and how to use it well.

Start here if AI still feels like magic or hype. In five short modules you will understand what today's AI systems really are, why they are so capable and so fallible, and how to use them safely in everyday work.

• Explain the difference between AI, machine learning and generative AI.
• Describe how a language model produces an answer.
• Spot the situations where AI is likely to be wrong.

Get great results on the first try, consistently.

Move from chatting to directing. This path teaches the structure behind every strong prompt, how to use context and examples, how to get output in exactly the format you need, and how to debug a prompt that is not working.

• Write prompts with a clear goal, context, constraints and format.
• Use examples to control tone and structure.
• Diagnose and fix prompts that give generic or wrong output.

Images and video that look intentional, not generated.

Learn how image and video models work, how to write prompts that give you control over composition, light and style, how to edit and keep characters consistent, and how to use these tools responsibly.

• Write image prompts that control subject, composition, lighting and style.
• Edit images precisely and keep characters consistent.
• Direct short video clips with shot language.

Pick the right model for every job, and know why.

There are dozens of capable models and new ones every month. This path teaches you to read the landscape, interpret benchmarks without being fooled, trade off cost, speed and quality, and choose between open and closed models.

• Map the major labs and model tiers.
• Read a benchmark result critically.
• Estimate the cost of a workload.

Design agents that finish the job safely.

Go from understanding the agent loop to building reliable agents: designing tools, engineering context, choosing an architecture, putting guardrails in place and measuring whether it actually works.

• Explain the agent loop and its failure points.
• Design tools and MCP servers a model can use well.
• Choose the simplest architecture that works.

Under the hood: transformers, RAG, evals, tuning and inference.

For engineers building production AI systems. Understand the transformer and mixture-of-experts architectures, build retrieval that works, run evals in production, decide when to fine-tune, optimise inference cost, and understand how labs make models safe.

• Explain attention, MoE and scaling at a technical level.
• Design and debug a RAG pipeline.
• Run evals and monitoring in production.