AI for the Curious

AI for the Curious · Course catalog

Every course, beginner to expert

The full path, broken into small courses you actually do. Click any course to open its episodes: the individual read → see → do units inside it. Every course is listed, including the ones still being built, so you can see the whole path. Expect courses and episodes to be added, split, or reworded as we build.

The first lessons are being built right now Last updated 2026‑09‑24

The first 5 lessons are open to everyone. Every other lesson is free too; it just needs a free account (email and password, or a sign-in link by email).

Read it top to bottom

Four levels, easy to hard, grouped into sections, then courses, then episodes.

Or follow one stage

Every course is tagged with a stage of the journey (Understand, Use, Build, Train, Evaluate, Govern). Filter to one to see just that lane, beginner to expert.

Filter by stage

·

No courses in this stage yet.

Level 1

Beginner

You have never really used AI, or only dabbled. By the end you can use it confidently and safely for real, everyday tasks. Almost everything here is in the Use and Understand stages.

First steps with AI

What AI Is (and Isn't)Sort real capabilities from myths by testing claims against live examples.Understand

Episodes

  1. What AI can and can't doSort a pile of claims into real versus hype.
  2. Why AI matters nowPlace the milestones that made it mainstream on a timeline.
  3. Myths vs realityTap each common misconception and see the correction.
  4. Capabilities and limitsMatch tasks to what AI does well or badly.
  5. Setting realistic expectationsAnswer a scenario checkpoint to finish.
The AI Landscape TourMatch tasks to the right kind of AI: chat, image, audio, video, code, and more.Understand

Episodes

  1. The map of AI typesGet the big picture of the main categories.
  2. Chat, reasoning and assistantsPick the right conversational tool for a question.
  3. Making things: image, video, audioMatch a creative task to the tool that fits.
  4. Words and knowledge: writing and researchChoose a writing or research AI for a goal.
  5. Getting things done: productivity, data, designRoute work tasks to the right kind of assistant.
  6. The frontier: automation, science, roboticsRecognize where AI is heading next.
Your First ConversationsRun a guided chat and practice steering, correcting, and following up.Use

Episodes

  1. Starting a chatSend your first prompts and read the reply.
  2. Steering the answerRedirect and refine in the middle of a chat.
  3. Following up and correctingIterate across turns to a better result.
  4. Saving a good chatKeep and reuse what worked.

Getting to know the major assistants

The Major AI AssistantsLearn what each popular assistant is best at, and how to move between them.Use

Episodes

  1. The main players todaySee the current lineup and who makes each.
  2. Using ClaudeIts strengths, standout features, and good first tasks.
  3. Using ChatGPTIts strengths, standout features, and good first tasks.
  4. Using GeminiIts strengths, standout features, and good first tasks.
  5. Using GrokIts strengths, standout features, and good first tasks.
  6. Using Microsoft CopilotIts strengths, and where it fits in daily work.
  7. Switching between toolsCarry the same skill from one assistant to another.

Getting productive

Prompting FundamentalsRewrite weak prompts into strong ones using context, constraints, and examples.Use

Episodes

  1. Anatomy of a promptLabel the parts of a strong prompt.
  2. Context and constraintsAdd the right background and limits.
  3. Examples and personasSteer output with samples and roles.
  4. Tone and formatControl the voice and shape of the answer.
  5. Iterating to better outputImprove one prompt across several turns.
Thinking With AIUse AI as a tutor, brainstorm partner, and reviewer through back-and-forth.Use

Episodes

  1. AI as collaboratorCo-work through a task together.
  2. AI as tutor and coachLearn a new topic with it.
  3. Learning without outsourcing the thinkingStudy with AI and still do the thinking yourself.
  4. AI as brainstorm partnerGenerate and expand ideas.
  5. AI as reviewerGet feedback and critique on your work.
  6. The iterative conversationBuild a result across a dialogue.
Everyday AIApply AI to real personal tasks: planning, writing, research, decisions.Use

Episodes

  1. Writing and readingDraft, summarize, and rephrase.
  2. Planning and researchPlan a project or a trip.
  3. Home life: cooking, fitness, financeEveryday helpers for daily life.
  4. Learning and hobbiesUse AI for personal growth and fun.
Choosing the Right ToolPick the best AI for a job, weighing cost, quality, privacy, and speed.Use

Episodes

  1. Match the tool to the taskPick by the kind of job.
  2. General vs specialistKnow when to reach for each.
  3. Cost vs qualityTrade speed, price, and results.
  4. Local vs cloud, open vs commercialWeigh privacy and control.

Using AI safely

Spotting Mistakes & HallucinationsCatch and verify wrong answers with a repeatable checking routine.Use

Episodes

  1. Why AI makes things upUnderstand what a hallucination is.
  2. Red flags in an answerSpot the signs of a likely error.
  3. A verification routineCheck claims quickly and reliably.
  4. When to trust, when to checkCalibrate how much to rely on an answer.
Privacy, Copyright & DeepfakesDecide what is safe to share and learn to recognize manipulated media.Understand

Episodes

  1. What not to shareProtect private and sensitive data.
  2. Who owns AI outputCopyright and ownership basics.
  3. Bias in AIRecognize skewed or unfair results.
  4. Spotting deepfakesDetect manipulated images, audio, and video.
Scam & Fraud AwarenessIdentify AI-powered scams and social engineering before they land.Use

Episodes

  1. AI-powered scams todayRecognize the new wave of threats.
  2. Voice and video impersonationVerify who you are really talking to.
  3. Prompt injection, for usersSee why pasted text can be risky.
  4. A personal safety checklistBuild habits that protect you.

How AI works (just enough)

How Models Read Text: TokensBreak text into tokens and predict token counts yourself.Understand

Episodes

  1. Words become tokensWatch text get split into pieces.
  2. Counting tokensPredict how many tokens a text becomes.
  3. Why tokens matterSee how they drive cost, limits, and quirks.
Context Windows & MemorySee why long chats forget, and manage what the model remembers.Understand

Episodes

  1. What the model can seeUnderstand the context window.
  2. Why long chats forgetWatch the window fill up.
  3. Managing contextKeep the right things in view.
  4. Extending memoryA first look at projects, notes, and retrieval.
How LLMs Learn, in Plain LanguageTrace training versus inference without touching the math.Understand

Episodes

  1. Prediction, not thinkingSee how a model generates text.
  2. Training vs inferenceTell the two phases apart.
  3. Where the knowledge comes fromTrace data and patterns to answers.
  4. Why they still get things wrongUnderstand the limits built into the design.
Level 1 examCertificate: AI Literate. 30 questions from lessons 1–14 and their exercises. Opens when every lesson of the level and its 3 picks are done; pass mark 80%.Open the exam →
Level 2

Intermediate

You are comfortable using AI. Now you go deeper at work, become a power user, and build your first simple things with AI. The Build stage starts here.

AI at work

AI for Your ProfessionBuild a job-specific workflow for your actual role and tasks.Use

Episodes

  1. Map your weekly tasksFind where AI can help most.
  2. Build a role workflowAssemble a daily assistant for your job.
  3. Draft, review, decideApply it to real work.
  4. Share with your teamStandardize what works.
AI Rules at WorkKnow what you can share, when to disclose AI help, and how to follow your workplace policy.Govern

Episodes

  1. Why workplaces have AI rulesSee what is at stake: data, trust, and the law.
  2. What you can and cannot shareSort real examples into safe and unsafe.
  3. Disclosure and creditDecide when to say that AI helped.
  4. Reading an AI policyFind the rules that apply to you.
  5. When there is no policyUse a simple personal rulebook.
Working With Documents & FilesExtract, summarize, and transform PDFs, spreadsheets, and contracts.Use

Episodes

  1. PDFs and long documentsExtract and summarize the key points.
  2. Spreadsheets and data filesAnalyze tabular data with AI.
  3. Contracts and reportsReview and flag what matters.
  4. Books and research papersDigest and question dense material.
Researching With AIRun a multi-step research task with citations, and check every source that matters.Use

Episodes

  1. Search versus research modesKnow when a quick answer is enough.
  2. Framing a research questionScope it so the answer is useful.
  3. Working over your own sourcesGround answers in documents you trust.
  4. Checking the citationsOpen the sources and confirm the claims.
  5. Writing up what you foundTurn findings into a summary you can defend.
Voice & Multimodal AIUse vision, screenshots, OCR, and voice to get work done hands-free.Use

Episodes

  1. Talking to AIHold a voice conversation.
  2. Show it your screenUse screenshots and vision.
  3. Reading the world: OCR and camerasCapture text and objects around you.
  4. Live translationCross a language gap in real time.
Creating Images, Audio & Video With AIMake images, voiceovers, and short clips, and know the rights and labels that come with them.Use

Episodes

  1. Prompting for imagesDescribe subject, style, and composition.
  2. Editing and iteratingFix details instead of starting over.
  3. Voice and musicGenerate narration and background audio.
  4. Short video clipsTurn a still or a script into motion.
  5. Rights, labels, and content credentialsKnow what you may use and how to label it.
Data & Spreadsheets With AIAnalyze and visualize a real dataset by directing an AI assistant.Use

Episodes

  1. Ask questions of your dataAnalyze in plain language.
  2. Clean and reshapeFix messy data with AI.
  3. Charts and summariesVisualize the results.
  4. Simple forecastsProject a trend forward.
AI Coding AssistantsUse the popular coding assistants to write, explain, and fix code faster.Use

Episodes

  1. What coding assistants doAutocomplete, chat, and agentic edits.
  2. GitHub CopilotIn-editor completions and chat.
  3. CursorAn AI-first code editor.
  4. Claude Code and CodexAgentic coding from the terminal.
  5. Choosing and combining themPick the right tool for the job.

Becoming a power user

Prompt Templates & LibrariesPackage your best prompts into reusable, shareable templates.Use

Episodes

  1. From good prompt to templateParameterize what you reuse.
  2. Build a personal libraryOrganize your prompts.
  3. Share and versionKeep templates current over time.
Building Skills & Custom WorkflowsCreate a repeatable multi-step skill the AI runs on command.Use

Episodes

  1. What a skill isUnderstand a saved multi-step routine.
  2. Design a skillBreak a task into clear steps.
  3. Build and test itMake it repeatable and reliable.
  4. Chain skills togetherCompose bigger workflows.
Projects, Memory & ContextSet up persistent projects so the AI keeps the right context.Use

Episodes

  1. Set up a projectGroup related work in one place.
  2. Give it lasting memoryAdd persistent context.
  3. Manage what it remembersPrune and update the memory.
  4. Keep humans in the loopVerify outputs before you trust them.
Multi-Model WorkflowsRoute each part of a task to the model that handles it best.Use

Episodes

  1. Different models, different strengthsKnow the lineup.
  2. Route tasks to the best modelSplit the work sensibly.
  3. Compare outputsPick the strongest result.
  4. Combine resultsStitch pieces into a final answer.
Judging AI OutputCheck AI work against a few test cases before you trust it with anything that matters.Evaluate

Episodes

  1. Why "looks good" is not enoughSee how polished answers hide errors.
  2. Your personal test setWrite three cases that really matter.
  3. A quick scoring checklistRate accuracy, completeness, and fit.
  4. Catching quiet errorsSpot confident mistakes in real outputs.
Automating Tasks Without CodeConnect AI to your email, calendar, and spreadsheets with free no-code tools.Use

Episodes

  1. What automation can take off your plateSpot repetitive tasks worth automating.
  2. Triggers, steps, and actionsRead and build a simple flow.
  3. Adding an AI stepSummarize, sort, or draft inside a flow.
  4. Testing before trustingRun it on safe examples first.
  5. Keeping it safeLimit access and keep a human check.
Delegating to AI AgentsHand an agent a real task, set its limits, and check what it actually did.Use

Episodes

  1. What an agent actually doesSee the plan, act, and check loop.
  2. Writing a good briefGive it a goal, limits, and a definition of done.
  3. Permissions and limitsDecide what it may touch.
  4. Checking the workReview its actions, not just the result.
  5. Hijacking and other risksSpot pages that try to take over an agent.
Running Open Models on Your Own ComputerInstall a free local model and decide when local beats the cloud.Use

Episodes

  1. Open versus hosted modelsWeigh privacy, cost, and quality.
  2. Will it run on my machine?Match model size to your memory.
  3. Installing a local modelGet a free app running in minutes.
  4. Chatting privatelyUse it with your own documents.
  5. When to go back to the cloudKnow the limits of small models.

Your first AI apps

Talking to Models Through an APISend your first API calls and read the structured responses.Build

Episodes

  1. From chat box to APIThe same model, now in code.
  2. Your first API callSend a request and get a reply.
  3. Reading the responseParse and use the output.
  4. Keys, limits and costsUse APIs responsibly.
Wrap an LLM in a Simple AppBuild a working chat, summarize, or Q&A app with a basic UI.Build

Episodes

  1. A minimal app shellSet up the project.
  2. Send a prompt, show a replyBuild the core loop.
  3. A basic chat UIPut an interface on it.
  4. Summarize and Q&A modesReuse the same pattern.
Structured Outputs & StreamingGet reliable JSON out of a model and stream responses to the screen.Build

Episodes

  1. Getting JSON out reliablyAsk for structured output.
  2. Schemas and validationEnforce a shape on the response.
  3. Streaming responsesShow text token by token.
  4. Handling errorsAdd basic retries and fallbacks.
Prompt Engineering for BuildersApply few-shot, chain-of-thought, and templating inside code.Build

Episodes

  1. Zero-shot and few-shot in codePut examples in the prompt.
  2. Chain-of-thoughtAsk the model to show its reasoning.
  3. Prompt templates in codeParameterize prompts programmatically.
  4. Evaluating promptsMeasure which version works.

Foundations for going deeper

Python for AIWrite the Python you need: data handling, functions, environments.Train

Episodes

  1. Python essentialsSyntax and core data types.
  2. Functions and objectsOrganize your code.
  3. Working with dataLists, dictionaries, and files.
  4. Environments and packagesSet up a clean workspace.
  5. Notebooks and testingA workflow that scales.
Data LiteracyCollect, clean, label, and judge the quality of a dataset.Understand

Episodes

  1. Where data comes fromCollection and sources.
  2. Cleaning and labelingPrepare a usable dataset.
  3. FeaturesTurn raw data into signals.
  4. Quality, bias and privacyJudge whether a dataset is sound.
Math You Actually NeedBuild intuition for vectors, probability, and gradients through visuals.Train

Episodes

  1. Vectors and matricesLinear algebra intuition, visually.
  2. Probability and statisticsReason under uncertainty.
  3. Calculus for learningSee what gradients mean.
  4. OptimizationHow a model actually improves.
Level 2 examCertificate: AI Power User. 30 questions from lessons 15–37 and their exercises. Opens when every lesson of the level and its 3 picks are done; pass mark 80%.Open the exam →
Level 3

Advanced

You build real AI applications and start to understand models from the inside. Retrieval, agents, evaluation, and the first serious machine learning all live here.

Retrieval & knowledge

Embeddings & Semantic SearchTurn text into vectors and build a working semantic search.Build

Episodes

  1. What an embedding isMeaning represented as vectors.
  2. Measuring similarityDistance and nearest matches.
  3. Vector databasesStore and query embeddings.
  4. Build a semantic searchPut it together end to end.
Building a RAG SystemAssemble chunking, retrieval, and grounding into Q&A over your own docs.Build

Episodes

  1. Why RAGGround answers in your own data.
  2. Chunking documentsSplit content for retrieval.
  3. Retrieve and injectFeed context to the model.
  4. Grounding and citationsCite the sources behind an answer.
  5. Put it togetherA working document Q&A.
Advanced RAGAdd hybrid search, reranking, and agentic multi-step retrieval.Build

Episodes

  1. Hybrid searchCombine keyword and vector search.
  2. RerankingOrder retrieved results better.
  3. Query reformulationRewrite the question for better hits.
  4. Agentic, multi-step retrievalIterate toward an answer.
  5. Knowledge graphsRetrieve over structured relationships.

Tools & agents

Tool Use & Function CallingGive a model tools and let it call them to take real actions.Build

Episodes

  1. Giving a model toolsThe core idea of tool use.
  2. Function callingDefine and expose tools.
  3. Executing safelyValidate what the model wants to run.
  4. Tools in a loopAct, observe, and repeat.
MCP Servers & IntegrationsBuild an MCP server that plugs your data and tools into any AI.Build

Episodes

  1. What MCP isA universal protocol for tools and data.
  2. Anatomy of an MCP serverResources, tools, and prompts.
  3. Build your first serverExpose your own data.
  4. Connect it to an assistantWire it up end to end.
  5. Security and permissionsIntegrate safely.
Building AgentsCreate an agent that plans, uses tools, and reflects on its work.Build

Episodes

  1. What makes an agentBeyond a single model call.
  2. The ReAct loopReason and act in turns.
  3. Planning and reflectionThink ahead and self-correct.
  4. MemoryShort-term and long-term.
  5. Build a working agentComplete a real task.
Multi-Agent SystemsOrchestrate several agents with roles, memory, and hand-offs.Build

Episodes

  1. Why multiple agentsDivide and conquer a problem.
  2. Roles and orchestrationCoordinate specialized agents.
  3. Communication and hand-offsPass work between them.
  4. Shared state and memoryKeep agents in sync.
  5. A multi-agent workflowBuild one that works.
Guardrails & Human-in-the-LoopAdd retries, guardrails, and checkpoints so agents fail safely.Build

Episodes

  1. Where agents go wrongKnow the common failure modes.
  2. GuardrailsConstrain what an agent can do.
  3. Retries and fallbacksRecover from failures gracefully.
  4. Human checkpointsAdd approvals and oversight.

Evaluating & securing AI

Testing & Evaluating LLM AppsBuild an eval set and score model outputs automatically.Evaluate

Episodes

  1. Why eval is hardNon-determinism and open outputs.
  2. Building an eval setCollect golden examples.
  3. Automated metricsScore outputs at scale.
  4. A/B testingCompare two versions fairly.
LLM-as-a-Judge & GroundednessUse models to grade models, and measure faithfulness to sources.Evaluate

Episodes

  1. Models grading modelsThe judge pattern.
  2. Writing judge rubricsDefine what good looks like.
  3. Groundedness and faithfulnessMeasure hallucination against sources.
  4. Human evaluationKnow when and how to use people.
Security for AI AppsProbe your own app for prompt injection and jailbreaks.Evaluate

Episodes

  1. Prompt injectionThe top threat to LLM apps.
  2. JailbreaksHow guardrails get bypassed.
  3. Secure tool executionSandbox what agents can run.
  4. Red-teaming your appAttack it yourself to find holes.
Tracing & Monitoring LLM AppsTrace every step of an AI app and catch problems once real people use it.Evaluate

Episodes

  1. Why logs are not enoughSee what a trace shows that logs miss.
  2. Instrumenting an appRecord prompts, tool calls, and costs.
  3. Reading a traceFind the step that went wrong.
  4. Monitoring in productionTrack quality, speed, and spend over time.
  5. Feeding evals from tracesTurn real failures into test cases.

Classical machine learning

Classical ML FoundationsTrain and compare regression, trees, and boosting on real data.Train

Episodes

  1. Learning from dataThe core ML mindset.
  2. RegressionPredict numbers.
  3. ClassificationTrees and logistic regression.
  4. Boosting and ensemblesCombine models for strength.
  5. Train and evaluateRun a full pipeline.
Clustering & DimensionalityFind hidden structure with clustering and PCA.Train

Episodes

  1. Unsupervised learningFinding patterns without labels.
  2. ClusteringGroup similar data points.
  3. PCAReduce dimensions.
  4. Anomaly detectionFind the odd ones out.
Recommenders & Time-SeriesBuild a recommender and forecast a time-series.Train

Episodes

  1. How recommenders workCollaborative filtering basics.
  2. Build a recommenderSuggest items to a user.
  3. Time-series basicsData that changes over time.
  4. ForecastingPredict what comes next.
Model Development & Error AnalysisTune, validate, and diagnose where a model goes wrong.Train

Episodes

  1. Metrics that matterMeasure the right thing.
  2. Cross-validationGet trustworthy scores.
  3. Bias and varianceUnderfitting and overfitting.
  4. Hyperparameter tuningImprove a model systematically.
  5. Error analysis and explainabilityUnderstand the mistakes.

Deep learning

Neural Networks From ScratchBuild and train an MLP, implementing backpropagation yourself.Train

Episodes

  1. The neuron and the layerThe building blocks.
  2. Forward passMake a prediction.
  3. BackpropagationLearn by gradient.
  4. Optimizers and training loopsMake it actually learn.
  5. RegularizationHelp it generalize.
Computer Vision With CNNsTrain a CNN to classify and detect objects in images.Train

Episodes

  1. How computers see imagesPixels and channels.
  2. ConvolutionsThe heart of a CNN.
  3. Build an image classifierTrain a CNN on real images.
  4. Detection and segmentationGo beyond a single label.
  5. Vision transformersThe modern approach.
Sequence Models & RNNsModel sequences with RNNs and LSTMs.Train

Episodes

  1. Data in sequenceText and time-ordered data.
  2. RNNsMemory across steps.
  3. LSTM and GRUHold on to longer context.
  4. Where RNNs fall shortThe road to transformers.
Transformers From ScratchImplement attention and build a small working transformer.Train

Episodes

  1. The attention ideaFocus on what matters.
  2. Self-attentionImplement it yourself.
  3. Building a transformer blockAssemble the parts.
  4. Tokenizers and embeddingsGet text into the model.
  5. Train a tiny transformerWatch it learn.
Generative Models: GANs & DiffusionGenerate images and understand how diffusion works inside.Train

Episodes

  1. Generative vs discriminativeA different goal.
  2. AutoencodersCompress and reconstruct.
  3. GANsGenerator versus discriminator.
  4. Diffusion modelsDenoising into an image.
  5. Generate your own imagesHands-on generation.
Level 3 examCertificate: AI Builder. 30 questions from lessons 38–58 and their exercises. Opens when every lesson of the level and its 3 picks are done; pass mark 80%.Open the exam →
Level 4

Expert

You train large models, run them in production, specialize, and lead. The deepest and most specialized work, plus the governance and leadership layer.

Training large models

Fine-Tuning & PEFTFine-tune a model with LoRA and QLoRA on a task you choose.Train

Episodes

  1. When to fine-tuneVersus prompting and RAG.
  2. Preparing a datasetFormat examples for training.
  3. Full fine-tuningThe basic approach.
  4. PEFT: LoRA and QLoRATune efficiently.
  5. Evaluate the resultConfirm it actually helped.
Instruction Tuning & AlignmentAlign a model to preferences with SFT and RLHF or DPO.Train

Episodes

  1. From base to assistantWhy instruction tuning matters.
  2. SFTSupervised fine-tuning.
  3. RLHFLearning from human feedback.
  4. DPO and alternativesSimpler ways to align.
  5. Measuring alignmentSafety and quality checks.
Pretraining & Scaling LawsUnderstand how foundation models are trained and scaled.Train

Episodes

  1. What pretraining doesLearning language from scratch.
  2. Data and tokenization at scaleBuilding the corpus.
  3. Scaling lawsWhy bigger tends to be better.
  4. Compute and costThe economics of training.
NLP & Speech SystemsBuild modern NLP pipelines plus speech recognition and text-to-speech.Train

Episodes

  1. Modern NLP pipelineText to insight.
  2. Embeddings and classic NLPWord2Vec through BERT.
  3. Speech recognitionAudio to text.
  4. Text-to-speechText to audio.

Running AI in production

Deployment & ServingShip a model behind an API with batching and streaming.Build

Episodes

  1. From notebook to servicePackage a model to run.
  2. Serving behind an APIExpose it to callers.
  3. Batch vs realtimePick the right mode.
  4. Scaling and streamingHandle real load.
MLOps: Pipelines & MonitoringBuild a pipeline with tracking, a registry, and drift monitoring.Build

Episodes

  1. The ML lifecycleWhat happens beyond training.
  2. PipelinesAutomate the flow.
  3. Experiment tracking and registryVersion everything.
  4. Monitoring and driftWatch models in production.
  5. RetrainingClose the loop.
Release Strategies & TestingRoll out safely with canary, champion/challenger, and rollback.Build

Episodes

  1. Testing ML systemsWhat to test and how.
  2. Blue/green and canaryRoll out gradually.
  3. Champion/challengerCompare models live.
  4. Rollback and skewRecover and detect problems.
Performance EngineeringSpeed up training and inference: parallelism, quantization, KV cache.Train

Episodes

  1. Where the time goesProfile before optimizing.
  2. GPUs and precisionAccelerate the compute.
  3. Distributed and parallel trainingScale across devices.
  4. Inference optimizationKV cache, batching, quantization.
Cost, Latency & ThroughputEngineer a deployment to hit real cost and latency targets.Build

Episodes

  1. The three leversCost, latency, and throughput.
  2. Measuring themEstablish honest baselines.
  3. Optimizing servingBatching and caching.
  4. Hitting a targetMeet an SLA in an exercise.

Specialize & research

Specialized DomainsApply AI in a chosen field: healthcare, finance, science, robotics, edge.Train

Episodes

  1. Pick a domainVision, speech, finance, health, or science.
  2. Domain data and constraintsWhat makes it different.
  3. Edge and robotics AIAI in the physical world.
  4. A domain mini-projectApply it end to end.
The Research FrontierExplore mixture-of-experts, efficient attention, causal inference, world models.Train

Episodes

  1. Reading researchPapers and moving trends.
  2. Efficient modelsMixture-of-experts and efficient attention.
  3. Beyond current LLMsCausal inference and world models.
  4. Self-improving and continual learningLook at the horizon.

Ethics, governance & leadership

Responsible AI: Bias & FairnessAudit a system for bias and document your mitigations.Govern

Episodes

  1. Where bias comes fromData and design choices.
  2. Measuring fairnessMetrics and their trade-offs.
  3. MitigationsReduce the harm.
  4. Documenting and auditingMake it accountable.
Governance, Law & ComplianceMap IP, copyright, data residency, and the main compliance frameworks.Govern

Episodes

  1. IP and copyrightWho owns what.
  2. Privacy and data residencyWhere data is allowed to live.
  3. Compliance frameworksThe major ones to know.
  4. Audit trailsProve what happened.
Leading AI in an OrganizationBuild an adoption and governance plan for a real team.Govern

Episodes

  1. Building an AI strategyDecide where to start.
  2. Adoption and changeBring people along.
  3. Governance and riskSet the guardrails.
  4. Measuring impactTrack ROI and outcomes.

Capstone & graduation

Capstone ProjectShip an end-to-end project in your track and add it to your portfolio.Evaluate

Episodes

  1. Choose your projectScope it to your track.
  2. Build itTake it end to end.
  3. Evaluate and documentProve that it works.
  4. Show it offPresent it in your portfolio and share it.
Level 4 examCertificate: AI Practitioner. 30 questions from lessons 59–73, including the capstone. Opens when every lesson of the level and its 3 picks are done; pass mark 80%.Open the exam → Final examCertificate: AI for the Curious Graduate. 40 questions across the whole path, including scenarios that combine several levels. Opens when all four level exams are passed.Open the exam →