Blog | Pragmatic AI Labs

AI Tooling: 20-Course Coursera Specialization from Foundation Models to Production

2026-04-17

Build and deploy production AI systems across the full stack — from generative AI fundamentals on AWS through deterministic agents, multi-modal programming, and serverless multi-model architectures. The flagship Pragmatic AI Labs specialization: 20 courses, covering the entire AI engineering lifecycle.

Enroll on Coursera →

What You Will Build

Production systems on Amazon Bedrock and SageMaker, deterministic agents in Rust and Deno, multi-modal pipelines that turn screenshots into code, MCP servers with provable contracts, AI-augmented CI pipelines, and a serverless multi-model SaaS capstone. Every course ends with a hands-on capstone you can share as a portfolio artifact.

Courses in This Specialization

Foundation Models and Bedrock (1–4)

  1. Generative AI and Foundation Models on AWS — Tokenization, RAG, Bedrock, llama.cpp, SageMaker Canvas.
  2. Intelligent Applications with Amazon Bedrock — Bedrock console, Claude, knowledge bases, agents.
  3. Prompt Architecture and NLP on Amazon Bedrock — Token lifecycle, prompt-as-code, chain-of-thought, Ollama bridge.
  4. AI Orchestration: From Local Models to Cloud — Prompt pyramid, caching, Ollama, llamafile, GPU Spot.

Enterprise AI and Security (5–8)

  1. Enterprise AIOps with Amazon Q Business — Q Business, CloudShell, cost control, RAG workflows.
  2. AI Security and Governance on AWS — Guardrails, CloudTrail, auth patterns, SageMaker Clarify, Rust.
  3. AI-Powered Analytics and Performance Engineering — Lambda, Rust, Amazon Q, CodeCatalyst, benchmarking.
  4. CLI Automation with Amazon Q and CloudShell — Q CLI, Docker, CDK, Lambda, ECR, IaC.

Agents, Debugging, and Multi-Modal (9–12)

  1. Deterministic LLM Programming — Code quality, AST analysis, technical debt, PMAT, Elo ratings.
  2. Agentic AI: Actor Models and Subagent Architecture — Actix, Rust, Go, Deno, supervision trees.
  3. AI Debugging and Test-Driven Fixes — AI debugging, TDD, logging, context gathering.
  4. Multi-Modal AI — Copilot, screenshot-to-code, Playwright, MCP.

Privacy, Pipelines, and MCP (13–16)

  1. Privacy-Conscious Development with AI Assistants — GitHub Advanced Security, Dependabot, Grype, secure prompting.
  2. AI-Powered Data Pipelines with Deno — Deno tasks, pre-commit hooks, quality gates.
  3. Building Deterministic MCP Agents — MCP, provable contracts, property testing, Kani BMC.
  4. Conversational Bot Architecture with Rust and Deno — Tokio, async runtime, Discord, Bedrock.

Production and Capstone (17–20)

  1. AI Code Review Automation with GitHub Actions — Actions, LLM prompting, GitHub Marketplace.
  2. LLM Security and Vulnerabilities — Prompt injection, model theft, plugin security.
  3. Build a Production SaaS Application with AI — API design, Docker, GitHub Pages, test harnesses.
  4. AI Tooling Capstone: Serverless Multi-Model Systems — Cargo Lambda, Bedrock routing, YAML prompts, production deployment.

Who This Is For

Companion GitHub Repo

github.com/paiml/ai-tooling — capstones, hero SVGs, and course-structure validator.