Gauntlet AI office atrium and gallery space

Gauntlet Night School

Learn to Build Production AI

Watch free AI engineering classes on demand. Explore the library and learn at your own pace.

Browse the class library

Notify me about new Night School sessions

Get an email when new Night School recordings are added to the library.

Night School On Demand

Catch up on every class.

Browse the full library — every past session, free and on demand.

The AI Engineering Interview (Part 2)

You got the interview — how do you prepare?

What companies look for in AI engineers today — evals, system design, and communication — and how to tell your story so it lands.

Transitioning Into AI Engineering (Part 1)

How do you break into AI engineering right now?

Why speed stopped being the signal, which skills are rising and falling, and how to stand out by building for the company you want.

Build a Neural Network From Scratch (Part 2)

How does a neural network actually learn?

Turn words into vectors, train a classifier, then build and train a neural network from scratch — loss, backpropagation, and learning rate.

Build a Neural Network From Scratch (Part 1)

What is a machine learning model, really?

Models as functions, why data beats algorithms, and where embeddings come from — the mental model before you build one.

The Tools We're Testing Right Now

Which AI tools actually earn a place in your workflow?

Grokbot agent swarms for QA and research, ChatPRD for fast product docs — and how to decide what stays.

QuizCaster: Shipping an AI-Generated Game

Can you ship a real game with no game-dev background and no art team?

A Gauntlet grad built an educational roguelike — art, audio, and animation all AI-generated — in about a month.

The Go-to-Market Engineer

What changes when marketing hires an engineer?

Embedding an AI engineer to build Bullseye, a self-serve CMS, and MCP-powered ad workflows.

MCP Factory: Build Any MCP Server Fast

Why rebuild the same MCP plumbing every time?

A plugin-based factory: implement fetch and format, and it standardizes the rest.

Why the Environment Beats the Lesson Plan

What drives the other 90% of your growth?

Doing, feedback, a tribe, and LLMs — the formal curriculum is only 10%.

Hermes, OpenClaw, and Managed Agents

How do you let agents do real work safely?

Governance, memory, and guardrails — why the platform matters more than the model.

Planning, Specs, and Decision Logs

What separates AI engineers from vibe coders?

Spec-driven development and decision logs that take code from prototype to production.

The AI-First Product Loop

Why do most AI builds go wrong?

Because someone asked the model to invent the plan. The 5 moves that ship.

Fine-Tuning LLMs: When to Use It (and When to Skip)

Fine-tuning or RAG — which do you need?

Most teams fine-tune when a better prompt would've worked.

AI System Design: Why It's Critical

Why does system design matter more now that AI writes code?

AI ships 1.7x more bugs — judgment is what separates juniors from operators.

Multi-Agent Systems: Production Infrastructure

What does it take to ship a multi-agent app?

Demos are easy; shipping is about managed state.

From the Front Lines

What actually changes after the program?

Three Gauntlet grads on planning over typing, fundamentals, and landing AI-first roles.

Agentic RAG Cookbook

Why do RAG systems give confident wrong answers?

It's a retrieval problem, not a prompt problem. Build one from scratch.

AI Coding Tools for Production Engineers

Which AI coding tool is best?

No single best tool — the 4-part framework production engineers use to choose.

What Hiring Managers Look For in AI Engineers

Is knowing the tools enough to get hired?

No — tools are table stakes. The proof points that land the job.

AI Code Review: From Vibe Checks to Real QA

How do you QA AI-generated code?

The quality gates that catch what vibe checks miss.

How to Build a Software Factory

How is a software factory different from vibe coding?

Turn AI from a demo machine into a repeatable pipeline that ships.

What you'll learn

Most engineers can vibe-code a working prototype. The harder — and far more valuable — skill is shipping production AI: systems that hold up under real users, load, integration, and security. That's what Night School teaches.

Each free class takes one piece of the production-AI craft and goes deep — AI system design, evals and real QA, retrieval and agentic RAG, multi-agent infrastructure, tool selection, and more — taught by the engineers who do it at Gauntlet. New classes drop every week, so the curriculum keeps growing with the field.

Come for one class or work through the whole library. Either way, you'll leave able to build AI that ships — not just demos.

Frequently asked questions

What is Gauntlet Night School?

A free library of recorded AI engineering classes taught by Gauntlet engineers. Explore practical lessons on building production AI, from agentic RAG to system design.

Is Night School free?

Yes. The recorded classes in the Night School library are free to watch on demand.

Who is Night School for?

Software engineers who want to add the production gear — building AI that ships to real users, not just prototypes.

When can I watch the classes?

Watch the recorded classes whenever it suits you. Browse the library and choose a topic.

Where can I find the recordings?

The class library on this page links to the available recordings and lesson details.

Ready to go all in?

Browse the class library Apply to Gauntlet