Instructor-led Preview
The classroom, on demand.
The same material, taught rather than read. A virtual instructor works through each module on video, with captions and a full transcript, so you can watch a lesson instead of reading one. Same course, same knowledge checks, same sources, and one record either way.
1 lesson filmed so far out of 40 written for the classroom. Every module is already available to read, and anything you finish now carries straight over when its lesson is published.
What AI Does—and Does Not Do
Build an accurate first mental model of generative AI as pattern-driven software with useful capabilities, observable limitations, and evidence requirements.
AI Systems and Useful Work
Recognize what an AI system is, where it can help, and where human judgment remains essential.
How Language Models Produce Responses
Build a practical mental model of training, inference, token-by-token generation, and uncertainty.
Context, Tokens, and Modalities
Understand what fits into a model request, what can be lost, and how text, images, audio, and files change the workflow.
Privacy, Safety, and Responsible Use
Protect people and data by matching AI use, controls, verification, and accountability to real consequences.
Prompt with Purpose
Turn a real need into a bounded, testable request that an AI system can attempt and a person can evaluate.
Prompt Anatomy and Success Criteria
Turn an ambiguous request into a clear, testable working agreement with an AI system.
Context and Evidence Construction
Build a small, trustworthy working set that helps an AI system answer from evidence instead of guesswork.
Examples and Output Contracts
Use representative examples and explicit output contracts to make responses easier to consume and validate.
Verification and Iterative Improvement
Check AI output with evidence, diagnose the kind of failure, and improve the workflow without hiding uncertainty.
Research with Evidence
Plan, execute, verify, and recover a bounded AI-assisted research workflow.
Writing and Transformation Workflow
Use AI to transform material while preserving meaning, voice, evidence, privacy, and reviewability.
Coding and Analysis Workflow
Use AI to change code or analyze data through scoped plans, executable checks, and reversible recovery.
Safe Tool-Use Workflow
Control what an AI system can read or change through explicit tools, permissions, approvals, and recovery.
Agents, Tools, and Guardrails
Recognize agentic behavior, choose the simplest sufficient design, and place enforceable controls around loops that can affect real systems.
AI Foundations Capstone
Plan, execute, verify, and explain a complete AI-assisted workflow with evidence that another person can review.
Design a Bounded Agent Loop
Turn a goal into an observable plan-act-check loop with explicit state, budgets, approvals, and stopping conditions.
Design Typed Tool Contracts and Action Controls
Give an agent narrow, validated interfaces and enforce authority, idempotency, verification, and recovery outside the model.
Engineer Context for Reliable Decisions
Select, structure, budget, and refresh the information an agent needs while preserving provenance and trust boundaries.
Design Safe Agent Memory Boundaries
Separate short-lived context from durable memory and govern what may be written, retrieved, corrected, expired, and deleted.
Understand MCP Architecture and Contracts
Model the host, client, server, lifecycle, capabilities, primitives, and tool contracts that let AI applications connect to external context and actions.
Secure MCP Trust Boundaries
Threat-model remote and local MCP integrations, constrain authority, validate tokens and servers, and keep untrusted content from steering consequential actions.
Choose Reliable Orchestration Patterns
Match deterministic workflows, model-directed routing, manager patterns, parallel work, and review loops to the uncertainty and control needs of a task.
Build Auditable Multi-Agent Handoffs
Transfer control, context, authority, and evidence between specialized agents without losing user intent or creating hidden permission expansion.
Evaluate Agent Behavior with Evidence
Build representative, adversarial, and regression evaluation sets that measure outcomes, trajectories, safety, cost, and latency before agent changes reach production.
Observe Agent Quality, Cost, and Risk
Instrument agent runs with privacy-preserving traces, metrics, logs, and quality signals that support diagnosis without turning telemetry into a sensitive-data archive.
Operate, Recover, and Improve Agent Systems
Classify failures across the full agent stack, contain impact, reconcile uncertain actions, recover safely, and turn incidents into verified improvements.
Reliable Agent Capstone: Design, Test, and Operate
Prove that an agent workflow can be understood, constrained, evaluated, observed, recovered, and handed off with evidence.
Choose a Deployment Shape and Operating Model
Choose among workstation, container, on-premises, edge, and cloud shapes from workload evidence, data boundaries, ownership, and recovery needs.
Model Identity, Licensing, and Provenance
Identify the exact model artifact, distinguish open weights from Open Source AI, and make a reviewable license and lineage decision before deployment.
Model Artifact Integrity and Safe Promotion
Acquire model artifacts into quarantine, verify identity and provenance, inspect executable surfaces, and promote only an immutable reviewed bundle.
Hardware, Runtime, and Capacity Planning
Select a compatible runtime and size compute, accelerator memory, host memory, storage, and concurrency from measurements of the exact serving build.
Model Serving and API Compatibility Contracts
Expose a bounded model endpoint with explicit health, identity, request, response, streaming, error, authentication, and compatibility contracts.
Endpoint Identity, Network, and Secrets Security
Protect a self-hosted model endpoint with explicit principals, least privilege, bounded networks, managed secrets, abuse controls, and auditable revocation.
Evaluate the Exact Serving Build
Gate a self-hosted endpoint on representative quality, safety, security, performance, and recovery evidence from the exact deployable build.
Observability, Cost, and Performance
Instrument model serving with privacy-conscious logs, metrics, traces, alerts, and cost evidence that connect user outcomes to exact model and runtime identities.
Scaling, Failure, and Capacity Controls
Scale self-hosted inference from measured demand while preserving quality, admission, failure isolation, availability, cost, and recovery evidence.
Model Update and Rollback Lifecycle
Plan, qualify, release, observe, and reverse model-serving changes without losing exact identity, compatibility, learner evidence, or a known-good recovery path.
Model Incident Response and Recovery
Detect, contain, diagnose, recover, communicate, and learn from incidents across self-hosted model artifacts, endpoints, infrastructure, data, policy, and operations.
Self-Hosted Model Operations Capstone
Deploy and defend a portable self-hosted model service whose identity, security, evaluation, capacity, telemetry, lifecycle, recovery, and approvals are proven by reviewable evidence.
Nothing is generated while you watch. Every lesson is produced and reviewed before it is published, then served as a fixed package.