Build Your Own AI Agent
Jai Bhagat
AI Educator for Business Leaders
Go from AI prototype to production by building your own agent
Your company wants AI agents but you're stuck at prototype. You don't know how to evaluate if your AI is working, or understand what's happening at runtime in the agent harness.
This course teaches you by doing. You'll build your own AI agent using Cursor and Claude Code (or free alternatives like opencode.ai). By building, you'll learn the foundations that transfer to ANY AI tool — evals, human-in-the-loop patterns, and why spikes beat sprints for AI work.
You'll leave with a working agent you built yourself, plus the skills to lead AI projects at your company.
What you’ll learn
Transform from "stuck at prototype" to confidently leading AI projects with hands-on skills that transfer to any AI tool
Set up your dev environment with industry-leading AI tools
Make your first agent call and see it work in real-time
Compare paid vs open source tools to understand tradeoffs
Understand what "evals" means and why it matters for production AI
Build an eval framework for your agent that catches issues before users do
Learn why agile/sprints are dead for AI work and what replaces them
Become the human-in-the-loop who works with many AIs and offers insights
Learn directly from Jai
Jai Bhagat
Your favorite teacher's favorite teacher
Who this course is for
Product Managers (3-10 years) at tech companies whose teams are building AI features but struggling to get from prototype to production
Designers at tech companies who want to understand how AI agents work so they can design better AI-powered experiences
Junior-to-mid level engineers wondering what's next for them now that AI writes all the code.
What's included
Live sessions
Learn directly from Jai Bhagat in a real-time, interactive format.
Maven Guarantee
Your purchase is backed by the Maven Guarantee.
Course syllabus
Week 1
Jan
12
Session 1: Foundations & Setup
Jan
15
Session 2: Build & Evaluate
Schedule
Live sessions
5-10 hrs / week
Mon, Jan 12
5:00 PM—6:30 PM (UTC)
Thu, Jan 15
5:00 PM—6:30 PM (UTC)
Build Your Own Agent
2 hrs / week
Get a real agent working on your machine, deployed to the internet and observe the agent's performance using advanced LLM observability + evaluation techniques.
Async content
1 hr / week
Each week has pre-reading materials and your personalized action plan.