The map of the AI Engineering Foundations skill path — what each chapter teaches, how they fit together, and where to start

Write prompts that work like programs — structured, testable, and consistently effective

Master the art and science of curating optimal context for AI agents

Constrained decoding, schema engineering, and type-safe AI pipelines

Build the minimum viable RAG pipeline — chunk, embed, store, retrieve, augment, generate — in plain code

Take the retrieval backbone to production — embeddings, chunking strategies, hybrid retrieval, advanced patterns, and evaluation. Assumes RAG Foundations.

When to fine-tune vs RAG vs prompting — SFT, LoRA, dataset prep, and evaluation

Vision, document, PDF, and video processing — build with images and audio, not just text

Build high-quality training and eval datasets — synthetic data, labeling, and curation pipelines

Prompt caching, KV-cache, batching, speculative decoding, and token budgeting

Architect knowledge bases that AI agents can navigate, retrieve from, and act upon

Pick the right post-training algorithm — DPO, GRPO, Dr-GRPO, DAPO, KTO, GiGPO — without drowning in research papers

Orientation across the five-chapter beginner path — why 2026 is the inflection point, why your soft skills are the moat, why you don't have to wait for technical specialists, what real companies are actually doing, and what to do this week

Why 2026 is the year AI quietly stopped being a side project for technical teams and became a daily tool for everyone else — and what doing nothing now actually costs

Judgment, communication, taste, domain expertise — the soft skills you've spent your career building are the moat. Walk through each one with concrete examples of it in action with AI

AI closes the gap between non-tech workers and the specialists you used to depend on — data analysts, designers, translators, legal reviewers. Concrete moves you can make without becoming a developer

Six to eight named companies, public numbers, and role-level stories — what Klarna, Shopify, Anthropic itself, and a marketer, salesperson, HR partner, and finance analyst actually did with AI in 2025–26

Pick one tool, three tasks, and a 4-question filter for when AI is the right call — the concrete plan you can start Monday morning, and the explicit handoff to AI Foundations as the next path

A 12-minute orientation to the Claude Code skill path — why it exists, what you'll be able to do, how the chapters relate, and where to begin

From zero to productive — master Claude Code fundamentals and start shipping with AI from day one

Advanced workflows, custom commands, and power-user patterns that 10x your output

Build your first Claude Code plugin — from zero to shareable

Advanced plugin patterns — hooks, agents, MCP servers, and marketplaces

Master every persistence layer — CLAUDE.md, auto memory, rules, settings, and hooks — so Claude never forgets

Tame your AI agent spend — model tiers, thinking budgets, caching, and Claude Code settings that cut costs without cutting quality

The orientation chapter — what makes Cowork its own discipline, and how the rest of the skill path is laid out.

Same model, three surfaces, three primary verbs — pick the right one for the task in front of you, and recognize the mid-task moment to switch.

The day-one chapter — install Cowork, connect three apps, and ship your first delegated deliverable from your filesystem before your coffee gets cold.

Three concentric rings of customization — Projects, Skills, Plugins — that turn the generic agent into a workspace shaped to one specific job.

Five power features — Dispatch, /schedule, Claude in Chrome, Live Artifacts, and sub-agents — that take you from productive Cowork user to running Cowork like an analyst who automated their job.

The operating manual for the one person at a Team or Enterprise tenant who configures Cowork org-wide — plans, identity, Memory policy, plugin marketplace, audit and compliance.

A 12-minute orientation across the five-app M365 Copilot deep-dive set — Outlook, Word, Excel, PowerPoint, Teams. One grammar (Agent Mode + License/Graph/Labels + verify-before-trust); five surfaces; no required order. Pick the app where your day actually lives

From 47 emails to 5 minutes — agent-mode triage, thread summarization that doesn't lose detail, on-brand drafting, natural-language rules, and the verification habits that keep Copilot honest in your inbox starting Tuesday morning

From blank page to polished doc with Agent Mode (default since April 22, 2026) — prompt-first drafting grounded in your files, brand-voice rewrites that don't sound like a robot, long-form structure that survives partial regeneration, and the four-check verification that keeps Copilot honest in your memos and proposals starting Tuesday morning

From a 2,000-row export to a board-ready answer in 15 minutes — Agent Mode (default since April 22, 2026), the brand-new =COPILOT() cell function, Python in Excel via natural-language prompts, the seven reliable data cleanups, and the four-check verification habits that keep your numbers honest in spreadsheets starting Tuesday morning

From a 9am brief to a 4pm board deck — Agent Mode (default since April 22, 2026), the Brand Kit + Template combo that fixes the 'why do my Copilot slides look generic' problem, exact-hex-and-template-name prompts, per-slide refinement that doesn't cascade, and the four-check verification habit that keeps your deck on-brand starting Tuesday morning

From eight hours of meetings to five minutes of catch-up — real-time mid-meeting questions, time-bounded prompts, action items with real owners (not fabricated), the Microsoft Tech Community v1.5 prompt for high-fidelity meeting analysis, and the four-check verification habit that keeps Copilot honest in your meetings starting Tuesday morning