How to Do AI Evals Step-by-Step with Real Production Data | Tutorial by Hamel Husain and Shreya Shankar
The Growth Podcast
Why application-specific evals matter
Hamel and Shreya explain why most AI features need systematic evals beyond demos and dogfooding.
Today’s Episode
Everyone’s demoing AI features. Few are shipping them to production reliably.
The gap? Evals.
Not the theoretical kind. The real-world kind that catches bugs before users do.
Hamel Husain and Shreya Shankar train people at OpenAI, Anthropic, Google, and Meta on how to build AI products that actually work. Their Maven course is the top-grossing course on the platform.
Today, they’re walking you through their complete eval process.
----
Brought to you by:
* The AI Evals Course for PMs & Engineers: You get $800 with this link
* Vanta: Automate compliance, Get $1,000 with my link
* Jira Product Discovery: Plan with purpose, ship with confidence
* Land PM job: 12-week experience to master getting a PM job
* Pendo: the #1 Software Experience Management Platform
----
If you want access to my AI tool stack - Dovetail, Arize, Linear, Descript, Reforge Build, DeepSky, Relay.app, Magic Patterns, and Mobbin - for free, grab Aakash’s bundle.
Are you searching for a PM job? Join me + 29 others for an intensive 12-week experience to master getting a PM job. Only 23 seats left.
----
Key Takeaways:
1. AI evals are the #1 most important new skill for PMs in 2025 - Even Claude Code teams do evals upstream. For custom applications, systematic evaluation is non-negotiable. Dog fooding alone isn't enough at scale.
2. Error analysis is the secret weapon most teams skip - Looking at 100 traces teaches you more than any generic metric. Hamel: "If you try to use helpfulness scores, the LLM won't catch the real product issues."
3. Use observability tools but don't depend on them completely - Brain Trust, LangSmith, Arise all work. But Shreya and Hamel teach students to vibe code their own trace viewers. Sometimes CSV files are enough to start.
4. Never use agreement as your eval metric - It's a trap. A judge that always says "pass" can have 90% accuracy if failures are rare. Use TPR (true positive rate) and TNR (true negative rate) instead.
5. Open coding then axial coding reveals patterns - Write notes on 100 traces without root cause analysis. Then categorize into 5-6 actionable themes. Use LLMs to help but refine manually.
6. Product managers must do the error analysis themselves - Don't outsource to developers. Engineers lack domain context. Hamel: "It's almost a tragedy to separate the prompt from the product manager because it's English."
7. Real traces reveal what demos hide - Chat GPT said the assistant was correct but missed: wrong bathroom configuration, markdown in SMS, double-booked tours, ignored handoff requests.
8. Binary scores beat 1-5 scales for LLM judges - Easier to validate alignment. Business decisions are binary anyway. LLMs struggle with nuanced numerical scoring.
9. Code-based evals for formatting, LLM judges for subjective calls - Markdown in text messages? Write a simple assertion. Human handoff quality? Need an LLM judge with proper rubric.
10. Start with traces even before launch - Dog food your own app. Recruit friends as beta testers. Generate synthetic inputs only as last resort. Error analysis works best with real user behavior.-
---
Where To Find Them
* LinkedIn:
* Hamel: Hamel’s LinkedIn
* Shreya: Shreya’s LinkedIn
* AI Evals Course: World’s best AI Evals Course (You get $800 off with this link)
Related Content
Newsletters:
* AI Evals
Podcasts:
* AI Evals: Everything You Need to Know to Start
* Everything you need to know about AI (for PMs and builders)
* Carl Vellotti on Claude Code
* Marily Nika on Google AI PM Tool Stack
PS. Please subscribe on YouTube and follow on Apple & Spotify. It helps!
This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.news.aakashg.com/subscribe
The AI-powered Podcast Player
