Demo agents are easy

Production agents get hired

Six weeks. One production-ready multi-agent system. Built, evaluated, and shipped to Google Cloud Platform. The same engineering patterns top teams use today.

Let's become a real Agent AI Engineer.

 

A 6-week course featuring live sessions with top AI experts Miguel Otero & Luis Serrano 

JOIN OUR WAITLIST

Sign up below to be the first to know when we offer the next course. You will receive information via email when enrollment opens. 

Thank you and we look forward to staying in touch!

 

Why a course about agents in production?

You already know what agents are. The problem is that most of them never leave the notebook.

Agentic AI is genuinely amazing, but only when backed by solid engineering principles. That's the gap this course closes. We start by grounding you in the concepts that actually underlie Agentic AI, then teach you how to ship those systems to production.

Concepts first. Workflows vs. agents, the agent loop, multi-agent orchestration, evaluation. Real intuition for what's happening inside the system, not just framework recipes.

Then, production. Containerization, IAM, observability, CI/CD, eval gates. The unglamorous parts that decide whether your agent runs once on your laptop or runs reliably on Google Cloud.

 By the end, you'll have what most engineers don't: 

A real, deployed, evaluated agent system.

 

Deploy a real agentic system

By the end of the course, you'll have shipped DocuAgent.

A multi-agent RAG system built with Google ADK and fully deployed to Google Cloud Platform (GCP).

PDFs flow through Cloud Storage and Pub/Sub into a modern OCR pipeline that captures text, tables, and figures as multimodal embeddings in Qdrant.

A Supervisor Agent then coordinates Retrieval and Web Search sub-agents over A2A, with Retrieval querying Qdrant via MCP.

Every agent runs on Cloud Run, with memory backed by Vertex AI, observability piped through OpenTelemetry into Cloud Trace and Cloud Logging, and credentials secured by IAM and Secret Manager.

Releases ship through an eval-gated CI/CD pipeline running Ragas, trajectory scoring, and LLM-as-judge.

A production-ready system you can put on your portfolio.

The kind that gets you hired.

What will I get?

6 live Q&A sessions

Each week you'll have the opportunity to ask all of your questions directly to our AI experts and learn from peers.

Certificate of Completion

After finishing the course you'll receive a Certificate from Serrano Academy and Neural Maze. 

Unlimited access for 6 months

You'll be able to go through the course at your own pace, with unlimited access to review all the materials for 6 months. 

Production-ready, multi-agent system

You'll have full access to code, labs, tools, to showcase and immediately use in your work.

Learning hub & community

You won't learn alone as you'll be invited to join a community-based learning environment from day one.

What will I build?

  • Know when to build a workflow vs. an agent Build the intuition for what agentic AI actually is, and when (and when not) to reach for it.
  • Speak the agent loop fluently Understand how agents perceive, reason, and act, and confidently debug where each step breaks down.
  • Design and orchestrate multi-agent systems Move from one overloaded agent to a coordinated team of specialized agents working seamlessly.
  • Build with industry protocols (MCP & A2A) Integrate the standards top teams are converging on, so your agents talk to tools and to each other the way the rest of the industry does.
  • Measure both RAG and agent quality Go beyond "vibes-based" testing with proper RAG metrics, trajectory scoring, and LLM-as-judge evaluation.
  • Ship to production on Google Cloud Cloud Run, IAM, Secret Manager, OpenTelemetry, and eval-gated CI/CD. The engineering work that turns a prototype into a system you can actually run.

Brought to you by

Faculty

Miguel Otero Pedrido

Senior AI / ML Engineer, physicist, and creator of the hugely popular The Neural Maze Substack and YouTube channel. Miguel has helped thousands of students learn how to build and ship real AI systems. He specializes in the part most courses skip: real infrastructure, real deployment, and real users. In this course, he owns the engineering: translating theory into working systems running on Google Cloud.

 

 

Luis Serrano

AI educator, mathematician, and founder of Serrano Academy, Luis has helped thousands of students around the world make sense of complex AI concepts through his popular courses, Grokking Machine Learning book, and YouTube channel. With a background at Google, Apple, and Cohere, he covers machine learning, NLP, deep learning, and quantum computing. In this course, he owns the theory: the intuition and mental models behind every concept you'll build.

 

Antonio Zarauz Moreno

R&D AI director and ML engineer dedicated to transforming research into reliable production systems. With expertise in LLM infrastructure, distributed training, and cloud deployment, he focuses on the practical bridge between prototypes and real-world operations. He brings a deep "builder’s" perspective to AI agents, specializing in how they connect to tools, scale, and survive failures. In this course, he owns the operations: the technical rigor and hard-won insights needed to move agents into a production-ready business environment.

You won’t just understand agentic AI.

You’ll build, test, and deploy a production-ready multi-agent system.

Who is this program for?

  • Students who want an edge: If you’re studying AI, data, or tech and want to go beyond theory, this course gives you hands-on experience building real systems you can showcase.

  • Professionals ready to level up: If you’re working in tech, marketing, product, or operations, learn how to apply agentic AI to automate workflows and solve complex problems in your field.

    Builders and creators: If you like turning ideas into reality, you’ll gain the skills to design and deploy AI systems that actually work—not just prototypes.

    Career switchers entering AI: If you’re transitioning into AI, this course helps you build practical, portfolio-ready projects that demonstrate real capability.

    Teams exploring AI adoption: If you’re evaluating how to bring AI into your organization, you’ll gain the frameworks and tools to make informed, strategic decisions.

     

    This Is NOT For You If

    • You’re only looking for theory without hands-on building
    • You want quick hacks instead of understanding how systems work
    • You’re not interested in working through real-world complexity

Frequently Asked Questions