Generative
AI.

F1 · Beginner · 30 min

A discriminative model tells cat from dog; a generative model paints a cat that never existed. Generative AI learns the underlying distribution of data — text, images, audio — and samples new examples from it. That one shift is what powers ChatGPT, Midjourney, and every model that creates.

• Generative vs discriminative
• Learning a distribution
• Sampling new data
• Modalities: text / image / audio

Build itList five products you use that generate content. For each, name the modality and whether it feels more like autocomplete or like painting from noise.

✓ CheckpointExplain what it means for a model to learn a distribution rather than a boundary, and why that is what lets it generate.

F2 · Beginner · 60 min

Text generation is next-token prediction on a loop: the model reads everything so far and picks the most likely next token, then repeats. Under it is the Transformer and self-attention. Understanding this makes prompting, streaming, and context limits all click.

• Next-token prediction
• Self-attention
• Autoregressive generation
• Streaming output

Build itGenerate the same prompt at temperature 0 and 1.0 several times. Explain, in terms of next-token probabilities, why one is repetitive and one is wild.

✓ CheckpointExplain why next-token prediction produces coherent paragraphs at all — the objective looks far too local for the result.

F3 · Beginner · 60 min

Image models work backwards from noise. In training they add noise to real images until they are static; at generation they start from pure noise and a trained network removes it step by step, sculpting a picture. That is the idea behind Stable Diffusion, DALL·E, and Midjourney.

• Forward noising
• Reverse denoising
• Sampling steps
• Text conditioning

Build itExplain to a friend why "add noise, then learn to remove it" is easier to train than generating a full image in one shot.

✓ CheckpointExplain what the reverse denoising process is actually undoing, and what the number of sampling steps buys you.

F4 · Intermediate · 45 min

To connect words and images you put both in one shared space, where a photo and its caption land close together. That is what CLIP does — and it is the bridge that lets a text prompt steer an image model. Multimodal generation lives on this idea.

• Shared embedding space
• Image–text alignment (CLIP)
• Cosine similarity
• Text-to-image guidance

Build itExplain how "a photo of a cat" can steer an image generator even though the model was never told what a cat looks like in labels.

✓ CheckpointExplain how an image and a caption end up close together in one embedding space, and what that enables.

F5 · Beginner · 45 min

A prompt is the steering wheel of a generative model. For text: roles, structure, examples. For images: subject, style, composition, and negative prompts. Learning to describe what you want — and what you do not — is the core creative skill of GenAI.

• Descriptive prompts
• Negative prompts
• Style & structure control
• Iteration

Build itTake one image idea and write three prompts: bare, detailed, and detailed-with-negatives. Compare how much control each gives.

✓ CheckpointExplain what a negative prompt is doing, and why it is not the same as omitting the thing.

F6 · Intermediate · 30 min

Generation is metered. Text is billed per token and bounded by a context window; images cost per generation and per resolution/steps. Knowing what drives cost and latency is what separates a fun demo from a product you can actually afford to run.

• Tokenization
• Context windows
• Cost per generation
• Latency drivers

Build itEstimate the cost of generating 1,000 product descriptions vs 1,000 images. Which is pricier, and what knob would you turn to cut it?

✓ CheckpointCompute the cost of one generation from tokens and price, then say which part of the prompt you would cut first.

T1 · Intermediate · 45 min

How a model picks the next token shapes everything. Temperature scales randomness, top-p and top-k prune the candidates, and beam search hunts for the most likely sequence. Tuning these trades creativity against reliability — the dial every text feature needs set right.

• Temperature
• Top-p / top-k
• Beam search
• Creativity vs reliability

Build itFor a legal-summary tool and a brainstorming tool, pick temperature and top-p for each and justify why they differ.

✓ CheckpointExplain what temperature and top-p each change about sampling, and why raising both at once is usually a mistake.

T2 · Intermediate · 60 min

Modern image models run diffusion in a compressed latent space for speed, guided by your prompt. Guidance scale controls how hard it follows the text; ControlNet-style conditioning adds pose, depth, or edges so you can direct composition, not just describe it.

• Latent diffusion
• Guidance scale
• ControlNet conditioning
• Img2img & inpainting

Build itExplain what raising the guidance scale does to an image, and when a lower value actually produces a better result.

✓ CheckpointExplain what guidance scale trades off, and what an extreme value does to the image.

T3 · Advanced · 60 min

Vision-language models take an image and text together and reason about both — describing a photo, answering questions about a chart, reading a screenshot. They fuse a vision encoder with a language model, and they are the fastest-moving frontier of GenAI.

• Vision-language models
• Image + text in
• Visual question answering
• Grounding to pixels

Build itList three product features that only become possible once a model can see an image and read text at the same time.

✓ CheckpointExplain what a vision-language model must align before it can answer a question about an image.

T4 · Advanced · 45 min

The same generative ideas extend to sound and motion: text-to-speech that sounds human, music from a prompt, and the young, fast-improving field of generated video. These push the hardest on compute, consistency over time, and evaluation.

• Text-to-speech
• Music generation
• Text-to-video
• Temporal consistency

Build itWhy is generating a 10-second video far harder than generating 10 separate images? Name the constraint that ties the frames together.

✓ CheckpointExplain why temporal coherence is the hard part of video generation rather than image quality.

T5 · Advanced · 60 min

To teach a model your style, brand, or a specific subject, fine-tune it. LoRA trains tiny adapters cheaply; DreamBooth teaches an image model a new subject from a handful of photos. The skill is knowing when a few example prompts would do instead.

• LoRA adapters
• DreamBooth
• Style vs subject tuning
• When not to fine-tune

Build itYou want a model to always draw your mascot. Argue whether prompting, LoRA, or DreamBooth fits — and what data each needs.

✓ CheckpointSay whether your problem is style or subject, and defend the tuning approach that follows.

T6 · Intermediate · 60 min

Left alone, generative models confidently make things up. Grounding them in retrieved sources — RAG — keeps generated text tied to real documents you control, so answers are checkable and citable. It is the main defence against hallucination in generative products.

• Retrieval-augmented generation
• Grounding
• Citations
• Hallucination control

Build itDesign a "chat with our help docs" feature. How do you make sure every answer is grounded in a real doc, with a link?

✓ CheckpointExplain why grounding reduces hallucination without eliminating it, and what a citation is really asserting.

P1 · Advanced · 90 min

How do you score something open-ended — a paragraph, an image? There is no single right answer. You combine human review, automated metrics, and LLM-as-judge into a repeatable eval so you can tell whether a change actually improved quality.

• Human evaluation
• Automated metrics
• LLM-as-judge
• Golden sets & regression

Build itDesign an eval for an image generator: what mix of human ratings and automated checks tells you a new model version is better?

✓ CheckpointExplain why an LLM judge that agrees with you 95% of the time may still be useless, and what you would check first.

P2 · Intermediate · 60 min

Generative models can be prompted into unsafe or off-brand output. Safety is layered: filter incoming prompts, moderate generated text and images, block disallowed content, and handle refusals gracefully — especially critical for anything public-facing.

• Prompt & output filtering
• Media moderation
• Disallowed content
• Refusal handling

Build itFor a public image generator, list the checks you run before showing a result to a user, and what happens when one trips.

✓ CheckpointExplain why output filtering is needed even when input filtering is in place.

P3 · Advanced · 60 min

Generation is slow and bursty. Streaming tokens as they are produced hides latency; batching improves throughput; and long generations or high-resolution images dominate cost. Serving GenAI well is about making the wait feel short and the bill stay small.

• Token streaming
• Batching
• Time-to-first-token
• Long-generation cost

Build itWhy does streaming make a 20-second answer feel fast? And what does batching buy you that streaming does not?

✓ CheckpointExplain why time-to-first-token matters more than total generation time for a streamed UI.

P4 · Intermediate · 45 min

Repeated or similar generations are everywhere — cache them. Route simple requests to smaller, cheaper models, cache identical prompts, and cap spend per user. In generative products, cost control is a feature users never see but the business always feels.

• Prompt / result caching
• Model routing
• Spend caps
• Cheap-model fallbacks

Build itDesign a cache for a text generator: when is a cached result safe to reuse, and when must you regenerate?

✓ CheckpointExplain why a semantic cache can serve a confidently wrong answer, and what threshold decision that risk lives behind.

P5 · Intermediate · 45 min

As generated media floods the world, proving what is AI-made — and where real media came from — matters. Watermarking embeds invisible signals in output; provenance standards (like content credentials) attach a tamper-evident history to a file.

• Invisible watermarking
• Content credentials
• Provenance metadata
• Detection limits

Build itWhy is watermarking generated text much harder than watermarking generated images? What can an attacker do to each?

✓ CheckpointExplain what a watermark can and cannot prove about a piece of generated media.

P6 · Advanced · 60 min

The finish line: fold everything — generation, grounding, evals, guardrails, serving, and cost control — into a real product people can use safely and you can afford to run. A conversational assistant is the canonical example that ties the whole roadmap together.

• End-to-end architecture
• Safety + cost + quality
• User experience
• Launch checklist

Build itDraw the architecture of a generative assistant: where do grounding, guardrails, caching, and evals each sit in the request path?

✓ CheckpointName the trade you made between safety, cost and quality in something you shipped — all three cannot be maximised.