---
title: "Learn"
description: "A free, open course on deterministic AI verification: why models cannot be trusted by default, how verification works, and how to build it into real systems."
url: https://qwedai.tech/learning
---

# Learn

A free, open course on making AI output provable: why language models cannot be trusted by default, how verification works, and how to build it into real systems. No GPU needed.

[Start the course](https://github.com/QWED-AI/qwed-learning) [Watch the 5-minute introduction](https://youtu.be/DjFOviJMMWY)

-   **17** modules and a capstone
-   **8 to 10 hours** for the full course
-   Free under **CC0-1.0**

## The model writes the draft. An exact check makes the decision.

Let the model do what it is good at, drafting and translating, and hand anything that must be exact to a deterministic checker. The course builds that habit from first principles to working code.

LLM **Draft** _Helpful, but may be wrong_

Checker **Decision** _Proven, or it stops_

## Choose your path

-   ### Backend developers

    About 90 minutes

    Add verification to the APIs you already run.

    Module 0 → Module 1 → Module 3 → Module 9

    [Start here](https://github.com/QWED-AI/qwed-learning/tree/main/module-0-prerequisites)
-   ### Finance and compliance

    About 2 hours

    Build checks for regulated work, from figures to legal claims.

    Module 1 → Module 2 → Module 6 → Module 11

    [Start here](https://github.com/QWED-AI/qwed-learning/tree/main/module-1-the-crisis)
-   ### AI and ML engineers

    The full course

    Every module, from theory to advanced patterns, and the capstone.

    All modules → Capstone

    [Start here](https://github.com/QWED-AI/qwed-learning/tree/main/module-0-prerequisites)
-   ### Students and researchers

    About 90 minutes

    The theory behind combining models with exact checks.

    Module 2 → Module 1.5 → Module 5 → Module 10

    [Start here](https://github.com/QWED-AI/qwed-learning/tree/main/module-2-neurosymbolic-theory)

## Course map

1.  [0 Prerequisites: trust-boundary basics 20 min](https://github.com/QWED-AI/qwed-learning/tree/main/module-0-prerequisites)
2.  [1 The crisis: why LLMs cannot be trusted by default 30 min](https://github.com/QWED-AI/qwed-learning/tree/main/module-1-the-crisis)
3.  [1.5 The physics of failure 45 min](https://github.com/QWED-AI/qwed-learning/tree/main/module-1.5-physics-of-failure)
4.  [2 Neurosymbolic theory: how verification works 45 min](https://github.com/QWED-AI/qwed-learning/tree/main/module-2-neurosymbolic-theory)
5.  [3 Hands-on: build your first verifier 60 min](https://github.com/QWED-AI/qwed-learning/tree/main/module-3-hands-on)
6.  [4 Advanced: patterns and integration 45 min](https://github.com/QWED-AI/qwed-learning/tree/main/module-4-advanced)
7.  [4.5 Production attestation lifecycle 60 min](https://github.com/QWED-AI/qwed-learning/tree/main/module-4.5-attestation-lifecycle)
8.  [5 The verification landscape 45 min](https://github.com/QWED-AI/qwed-learning/tree/main/module-5-verification-landscape)
9.  [6 Domain-specific verification 60 min](https://github.com/QWED-AI/qwed-learning/tree/main/module-6-domains)
10.  [7 Context engineering and its limits 60 min](https://github.com/QWED-AI/qwed-learning/tree/main/module-7-context-engineering)
11.  [8 Agentic integration 60 min](https://github.com/QWED-AI/qwed-learning/tree/main/module-8-agentic-workflows)
12.  [9 DevSecOps: CI/CD verification 45 min](https://github.com/QWED-AI/qwed-learning/tree/main/module-9-devsecops)
13.  [10 Advanced patterns: RAG and consensus 45 min](https://github.com/QWED-AI/qwed-learning/tree/main/module-10-advanced-patterns)
14.  [11 The legal auditor 60 min](https://github.com/QWED-AI/qwed-learning/tree/main/module-11-legal-auditor)
15.  [12 Agentic security guards 60 min](https://github.com/QWED-AI/qwed-learning/tree/main/module-12-agentic-security)
16.  [13 Secure agent orchestration 75 min](https://github.com/QWED-AI/qwed-learning/tree/main/module-13-secure-orchestration)
17.  [14 Infrastructure verification 75 min](https://github.com/QWED-AI/qwed-learning/tree/main/module-14-infrastructure-verification)
18.  [Capstone: a governed banking agent Project](https://github.com/QWED-AI/qwed-learning/tree/main/capstone-project)

Found a mistake or want to add a module? The course takes issues and pull requests like any other QWED repository: [QWED-AI/qwed-learning](https://github.com/QWED-AI/qwed-learning).