AI ENGINEER / 180
180 days · 2 hours a day · zero assumed knowledge

Beginner to job-ready
AI Engineer & Forward-Deployed Engineer

Not a link list. A guided, sequenced bootcamp: every day has an analogy-first explanation you can flip to full technical depth, code that runs in your browser, a step-through visualizer where a picture beats prose, a shipped artifact, a quiz that tracks mastery, and flashcards on a spaced-repetition schedule. It ends with a production RAG capstone, customer-simulation training, and interview prep.

Begin Day 1Open the dashboardSee all 180 days
What's inside
145
deep lessons
26
review checkpoints
9
dedicated project days
1
production capstone
540+
quiz questions
900+
flashcards, spaced
plus an AI teacher with explain / hint / mock-interview / customer-sim modes

The arc — nine phases

Phase 1 · days 1–21Code FoundationsPython from zero to fluent, plus the software-engineering habits — git, shell, testing, packaging, clean code — that everything else stands on.Phase 2 · days 22–49CS CoreData structures, algorithms, SQL, operating systems, networking, concurrency, security, and system design — the invisible machinery under every production system.Phase 3 · days 50–63Math for MLVectors, matrices, gradients, probability, statistics and entropy — exactly the math modern ML runs on, taught visually and by building.Phase 4 · days 64–84Data & Machine LearningFrom messy CSV to validated model: NumPy, pandas, EDA, feature engineering, supervised and unsupervised learning, metrics, and honest error analysis.Phase 5 · days 85–105Deep Learning & TransformersNeural networks from scratch, PyTorch, embeddings, attention, tokenization, and a tiny GPT you train yourself — the internals behind the APIs.Phase 6 · days 106–133AI EngineeringThe 2026 job core: LLM APIs, prompting, structured outputs, tool use, RAG, vector search, agents, MCP, fine-tuning, multimodal, and AI security.Phase 7 · days 134–147Evaluation & ObservabilityThe most underrated hiring signal: golden sets, LLM-as-judge, RAG metrics, regression gates, tracing, red teaming, and continuous evaluation.Phase 8 · days 148–161Production & MLOpsDocker, cloud, CI/CD, serving and inference optimization, monitoring, reliability, incident response — shipping AI that stays up and earns its cost.Phase 9 · days 162–180FDE & CapstoneForward-deployed craft: discovery, specs, prototypes, enterprise integration, stakeholder communication — then ship and present your capstone.

How every day works

step 1

The analogy

Every concept lands as a picture first — arrays are bookshelves, RAG is an open-book exam, evals are the exam you write before the student exists.

step 2

Flip to Tech

One toggle swaps the whole explanation for the precise version: invariants, complexity, trade-offs, correct terminology.

step 3

Watch it happen

Step-through visualizers animate the algorithm, the pipeline, or the attack — play, pause, and predict what happens next.

step 4

Run the code

Guided exercises execute right on the page (Python via WebAssembly). Edit, break, fix, re-run.

step 5

Ship something

Every day ends with an artifact against a measurable rubric — files, repos, services. Your portfolio builds itself.

step 6

Prove it & retain it

A 3-question mastery quiz routes misses to the exact day to revisit; flashcards return on a spaced-repetition schedule.

Built for the 2026 job market

The curriculum is weighted the way hiring is: RAG, agents, tool use, and evaluation get four full weeks; production engineering (Docker, CI/CD, serving, observability, cost) gets two; and the forward-deployed skillset — customer discovery, specs from ambiguity, prototypes, demos, enterprise integration — gets two more, with rubric-graded customer simulations. Interview training runs throughout and culminates in a full mock loop. Every resource is legally free and verified.