Class is in session

From Vibe to Live: Build and Deploy Production AI Agents

·

4 Weeks

·

Cohort-based Course

Build and deploy agentic apps in production from day one with Azure, OpenAI Agents SDK, Docker, FastAPI and Phoenix Arize.

Join students from

Meta
NVIDIA
Visa Inc.
Nubank
TD Bank

Course overview

Design, build and deploy production grade AI Agents from day one

Most courses stop at building toy notebooks. Here, you’ll learn to architect, evaluate, and deploy AI agents that solve real business challenges-factoring in security, scalability and operational costs. You’ll master not just the “how”, but the “when” and “why” behind agentic AI, with a special focus on software engineering, observability and scalability best practices.


Production-Ready, Not Just Toy Projects


Starting on the first day of this course, this is what you will work on:


✅ Building agentic apps with OpenAI Agents SDK

✅ Building observability pipelines with OpenTelemetry and Phoenix Arize to track what your agents are doing and understand why

✅ Containerizing and deploying your agentic application using Docker and FastAPI

✅ Implementing robust guardrails, so that responses from your agents are relevant and not prone to hallucination or prompt injection attacks


✨ Biggest differentiators of this course


Hands-On Product Development

From day one, you’ll build a real-world agentic product, guided by step-by-step tutorials. You’ll gain practical experience that goes far beyond theory


Free OpenAI Credits

Get USD 400 in free GPT 4.1 credits as soon as you enroll into this course


Individual Coaching Sessions

Feel free to schedule 30 minute individual sessions with me every week. You will get direct, personalised help, making sure you never lose momentum.


Industry Guest Lectures & Expert Workshops

Learn directly from the people shaping the future of AI. This cohort features guest lectures and workshops with industry experts, bringing you what's newest in AI Agent features for developing, evaluating and deploying AI Agents.


Capstone Project: From Idea to Production

You’ll connect best practices with hands-on implementation. For your capstone, you’ll design, evaluate, and deploy a production-grade AI agent. You will showcase your ability to bridge the gap between theory and enterprise impact.


Is this course for me?


👉🏻 If you have ever struggled with any of the following challenges:


🔹 Prototype Graveyard: You have built simple AI demos that never made it to production

🔹 Evaluation Paralysis: You struggle to measure if your agents are actually performing well

🔹 Security Vulnerabilities: You need to mitigate the risk that your agents are exposed to prompt injection attacks

🔹 Availability & Resiliency: You want to prevent agentic AI products from breaking after being deployed


Then yes, this course is for you.


🚀 What you will have built by the end of this course


✅ An end-to-end multiagent RAG system with clear agent scopes, tool calls and guardrails

✅ A production-deployed AI agent with a public URL you can share with employers

✅ Comprehensive observability pipelines to keep track of your agent's metrics and availability

✅ Portfolio-ready project demonstrating end-to-end agent development capability


🚀 Tools you will use


🦾 OpenAI Agents SDK for robust multiagent application development and orchestration

⚡️ Redis for lightning fast agentic memory implementation

📊 Phoenix Arize for end-to-end agent observability and evaluation

⚙️ Github Actions for seamless CICD pipeline orchestration

🐳 Docker for containerization and reproducible development

FastAPI for rapid Python backend development


🟢 Who this course is for


1. Software Engineers, Machine Learning Engineers, and Data Scientists

You have solid experience with data pipelines and classical ML problems. You’re ready to integrate large language models into agent architectures, build workflows using tool integrations, memory management, and task planning for practical AI systems.


2. Product Managers and AI Strategists

You translate business objectives into technical solutions. You want to understand how to define AI agent functionalities from product requirements, establish meaningful success metrics, and effectively collaborate with engineering teams to deliver scalable, user-centric agent applications.


3. Tech Leads and Backend Developers

You oversee complex systems involving multiple AI agents and services. You seek to master best practices for orchestrating, monitoring, and scaling agent workflows that run reliably in production environments.


🔴 Who this course is not for


1. Absolute Beginners in Programming or Software Development

This course assumes proficiency in Python and coding fundamentals. If you’re new to programming, start with foundational development courses before progressing here.


2. People looking for research heavy content

This practical course leverages pre-built language models and focuses on implementation, not on transformer internals, model training, or research heavy topics.


3. No-Code or Low-Code Developers

This program is code-centric, involving APIs and orchestration through programming libraries. If you prefer drag-and-drop or visual tools, this course may be more technical than expected


💸 Course Value vs Investment

Total Value: $3200 vs Investment: $1000


🔷 Live and on-demand training to help you design, build and deploy real AI Agent Apps

($1000 value)


🔷 Personalized support via Slack plus weekly 1x1 sessions (30 min x 4)

($800 value)


🔷 Free OpenAI Credits

($400 value)


🔷 Complimentary Access to All Future Cohorts

($1000 value)


Student Testimonials


🚀 "I truly appreciate our trainer Rafael. He has outstanding experience in AI, and I’m really glad to have him as my mentor. I’ve learned so much from his training sessions and the 1:1 online guidance. He clarified my doubts and explained how to build robust agentic AI systems with best practices something you normally don’t get from other courses. I’m definitely going to enroll in the upcoming programs. I highly recommend this program to anyone who wants to gain broader and deeper knowledge in AI."


🚀 "Glad to be in the first cohort! Agents development, functional tasks, openai SDK, evaluation techniques, observability and much much more. Thank you Rafa!"


🚀 "Got good understanding on how to build agentic apps. Best part of this course is the observability platform."


Ready to go beyond the PoC? Join us and build AI agents that matter-powered by real-world tools, enterprise best practices, and the support you need to succeed.

Who is this course for

01

Software Engineers, MLEs, and Data Scientists who want to build agentic apps using function calling, short term memory and observability

02

POs and PMs who want to dive deep into Agentic System architecture to deliver user-centric agentic apps in collaboration with engineers

03

Consultants, Solution Architects and Field Engineers who want to learn what it takes to successfully run agentic apps in production

Prerequisites

  • Basic Python knowledge

    This course is highly hands-on, so basic Python knowledge is required.

  • Familiarity with AI & Gen AI

    You use AI and Gen AI on a daily basis, and/or you are familiar with the main use cases surrounding it

  • Willingness to code full fledged products

    No notebook demos! You'll code, move fast and break stuff.

What you’ll get out of this course

Build an AI Agent on Day One

Theory is important: you will learn foundational concepts around AI Agents, but you will also start coding and shipping Agentic AI products from the first day.

Personalized Guidance

You will have access to a private Slack workspace where you can ask questions to me - or to fellow students. Want to chat, code together or brainstorm? Feel free to book a weekly 30 min 1x1 with me.

Free GPT Credits

You will get USD 400 in free OpenAI GPT 4.1 credits to build your agents. These credits are yours to use even after you finish the course.

Master and Implement AI Agent Evaluation

Evaluating AI Agents can be overwhelming. We will show you what to evaluate, how to do it and what to do with these metrics.

Understand where Agentic RAG shines, and use this knowledge to design and build unique AI products

Should I go for vanilla RAG? Should I implement Agentic RAG instead? You will deep dive into the main use cases for Agentic RAG, what problem it solves, when it should be used, and when it shouldn't.

AgentOps: Apply AI security and observability best practices

Learn about the best practices around tracking LLM calls, tools and agent interactions. Master prompt injection attacks and how to mitigate them. Incorporate all these concepts in your app.

Deploy your Agentic AI App Into Production and showcase your skills

Set up cloud resources, add authentication, deploy tracing and guardrails, release your app’s code and share your live project with the world

Guest Lectures from Industry Experts

You will learn from some of the best engineers and product professionals in the industry. In previous cohorts, we had guest speakers from Microsoft, LlamaIndex and Replit - stay tuned for the full speaker lineup for this cohort!

What’s included

Rafael Pierre

Live sessions

Learn directly from Rafael Pierre in a real-time, interactive format.

Free OpenAI Credits

USD 400 in GPT 4.1 credits that you can use for building and evaluating your agents

Guest lectures from experts in the field

Build with the latest & greatest insights from experts from Microsoft and Weights & Biases

Personal coaching with 1x1 sessions

Personalised help to make sure you don't get stuck

Community of peers

Stay accountable and share insights with like-minded professionals.

Lifetime access

Go back to course content and recordings whenever you need to.

Certificate of completion

Share your new skills with your employer or on LinkedIn.

Maven Guarantee

This course is backed by the Maven Guarantee. Students are eligible for a full refund up until the halfway point of the course.

Course syllabus

5 live sessions • 49 lessons • 3 projects

Week 1

Oct 3—Oct 5

    Oct

    3

    Live Session: Week 1

    Fri 10/35:00 PM—6:30 PM (UTC)

    Getting Started with Week 1

    2 items

    Agentic AI Applications & Agentic RAG

    3 items

    Optional: Evaluating Agentic AI Feasibility

    2 items

    Assignment: Version 1.0

    2 items

Week 2

Oct 6—Oct 12

    Oct

    9

    Optional: Guest Lecture 1

    Thu 10/95:00 PM—6:00 PM (UTC)
    Optional

    Oct

    10

    Live Session: Week 2

    Fri 10/105:00 PM—6:00 PM (UTC)

    Expanding Agent Capabilities: LLMOps & Observability

    7 items

    LLM as a Judge

    6 items

    Trajectory Evaluation

    7 items

    Assignment: Version 2.0

    1 item

Week 3

Oct 13—Oct 19

    Oct

    17

    Live Session: Week 3

    Fri 10/175:00 PM—6:30 PM (UTC)

    Contanerizing Your Application & Adding MCP Capabilities

    1 item

    Your First MCP Server

    4 items

    Assignment: Version 3.0

    6 items

    Bonus: Hosting MCP Servers on Azure

    2 items

Week 4

Oct 20—Oct 24

    Oct

    24

    Live Session: Week 4

    Fri 10/245:00 PM—6:00 PM (UTC)

    Bonus: Agentic AI & MCP Attacks & Mitigation Strategies

    4 items

    Production Deployment & Project Showcase

    4 items

    Bonus: Vibe Coding Your Agentic Frontend with Replit

    1 item

What students are saying

Success stories

        Not another POC, not just theory, but a true deep dive into what it takes to build production grade AI Agents. I came in thinking of agents as exciting prototypes. But, walking away knowing how to treat them as enterprise solutions. Observable, secure, and reliable.
Vinayak Talikot

Vinayak Talikot

Principal Software Engineer, Tech Lead
        I can confidently say it’s one of the most practical and value-packed programs I’ve attended. Rafael’s teaching style is clear, structured, and hands-on, making even complex concepts around production-grade AI agents easy to grasp. If you’re serious about mastering production AI agents this course is absolutely worth it.
Umamaheswara Pragada

Umamaheswara Pragada

Enterprise Solution Architect

Meet your instructor

Rafael Pierre

Rafael Pierre

AI Engineer with 17+ years experience - ex-Hugging Face, Databricks

Rafael is an AI expert with deep experience helping companies - from Fortune 500s to startups - architect, fine-tune, and deploy Large Language Models and Generative AI solutions.


With a background spanning software engineering, cloud computing, data & AI engineering, and solution architecture, he brings a full-stack, hands-on approach to building production-ready AI systems that follow best practices in MLOps and LLMOps.

Previously at

ABN AMRO
Hugging Face
O'Reilly Media
Databricks
ING
A pattern of wavy dots

Be the first to know about upcoming cohorts

From Vibe to Live: Build and Deploy Production AI Agents

Course schedule

4-6 hours per week

  • Fridays

    10 AM PST | 2 PM EST | 7 PM CET

    Recorded videos, live lessons, office hours and hands-on projects

Build Your First Agentic AI App with MCP

Build Your First Agentic AI App with MCP

Why MCP matters now

Learn the business case and technical advantages of MCP for modern AI apps

Map MCP architecture

Connect the components of MCP to real-world applications you can build

Build live, in real-time

Create your first working MCP-powered AI application step by step

Get the free recording

Learning is better with cohorts

Learning is better with cohorts

Building the Future Together, Live in Session

Self paced content, code examples and live sessions - multiple ways of learning. We'll go together.

Unleashing Agent Power: Build Powerful Agentic Apps with OpenAI Agents SDK

Go from first prototype to launching robust AI agents with practical insights and hands-on guidance

Design, Build, and Deploy Agentic Patterns in Action

Design and build workflows and dynamic agents. Choose the right pattern for each

challenge

Frequently Asked Questions

Stay in the loop

Sign up to be the first to know about course updates.

A pattern of wavy dots

Be the first to know about upcoming cohorts

From Vibe to Live: Build and Deploy Production AI Agents