Building Agents with Self-Learning Loops

Hosted by Mahesh Yadav

Fri, Apr 24, 2026

4:00 PM UTC (45 minutes)

Virtual (Zoom)

Free to join

4,808 students

Invite your network

Go deeper with a course

Agentic AI Product Management Certification
Mahesh Yadav
View syllabus

What you'll learn

Self-Improving Loops Every AI Builder Should Understand

Compare key loop patterns, where they work in production, and how to map them to real product use cases.

The 4-Part System Design for Self-Learning Agents

Design agents using signal, eval, memory, and adaptation with clear examples across product and engineering.

From Static Prompts to Adaptive Systems (Wave 1 → Wave 3)

Move beyond prompts to feedback-driven systems using memory and iteration, not fragile long context windows.

Failure Modes in Feedback Loops (and How to Avoid Them)

Spot common loop design failures early and apply guardrails to prevent cost, drift, and unreliable outputs.

Why this topic matters

For 70+ years, programming meant humans writing code & machines executing it. Autoresearch, Ralph, & Hermes invert this: humans write natural-language instructions (program.md, prd.json, SKILL.md), & the agent writes, modifies, and iterates on code. This isn't vibe coding — it's a new job description. The PM becomes a "programmer of programs.md" — authoring research org instructions, not Python.

You'll learn from

Mahesh Yadav

Ex AI Product Lead -Google l Meta l Microsoft l AWS | 10k+ Alums l Founder - Agentic AI Institute

Mahesh Yadav brings 20+ years of experience building AI products at Google, Meta, AWS, Microsoft. He holds 12 patents in AI training, power management and computer vision. He has launched major agentic-AI initiatives (for example launching an agent for AWS Bedrock, featured in CEO keynote) and trained thousands of professionals to succeed in AI roles. With this programme you get rare access to CEO-level of mentorship.
Substack AI PM Newsletter l Linkedin Community of AI PMs l YouTube For Free Sessions Recordings
Currently, he is building back-office AI agents for the enterprise, starting with in-house legal teams through LegalGraph.AI. His work bridges education and real-world AI deployment, helping organizations adopt agentic systems that automate complex knowledge work.

See all products from Mahesh

Previously at

Google
Microsoft
Amazon Web Services
Meta

Sign up to join this lesson

By continuing, you agree to Maven's Terms and Privacy Policy.