Maneesh Maddala

Principal Software Engineer in Testing

He/Him

Principal Software Engineer in Testing who loves to break things (in the best way) | Crafting reliable Modern Web Based Frameworks | Always testing, always learning—with a few laughs along the way.

🎂 MoTaBirthday | January 25, 2024

Open To

Mentor

CV Reviews

Teach

Speak

Moments (12)

Guardrails and evals are not scary!

Listening to Maneesh talk about guardrails and evals at the AI Chapter event. My insight is that a guardrail is a ... text, and evals are like a JSON that validates a JSON schema. 

23 Sep

Guardrails and evals people can trust

In the session "Guardrails & Evals: Keeping Your AI Skills Honest"  from  Maneesh Maddala in the AI Chapter Writing Guardrails:1. Make it testable: "A rule without a test is a sugges...

23 Sep

The eval pyramid

Where there is quality thinking there is a pyramid.In the session "Guardrails & Evals: Keeping Your AI Skills Honest"  from  Maneesh Maddala in the AI Chapter , the eval pyramid was s...

23 Sep

AI Chapter is coming up with more exciting events

Leading the MoTaverse AI Chapter comes with one perk. I get to invite the people I'd pay to learn from. And make it free for you. Here's who said yes for the next five months: 📅 23 Sep: Maneesh Mad...

19 Sep

Sessions (11)

Guardrails & Evals: Keeping your AI skills honest

Understand the necessity of implementing guardrails and evaluations when using AI tools in professional workflows.

23 Sep

The search for the lost code, best chair and the magic of MoTaCon

A lively episode about MoTaverse profile upgrades, MoTaCon momentum, AI-assisted testing, hidden knowledge silos, and why real human conversations still matter when tools move fast.

7 Aug

Glossary Terms (3)

Guardrails What it is:The safety rule. It must hold every time.From the repo:Never call a change "safe to ship" AI Evals The test for that rule.Does the assistant refuse to call it safe? Trial One run of the test. A pass only proves that one run.

Comments (2)

The eval pyramid Thanks for the adding this to the moment Rosie. We should treat evals exactly like we approach testing. After all, we are testers. Mapping the Machine: What Fell Out of an LLM Wiki Idea There’s already a skill/plugin for this. zits called graphify. https://github.com/Graphify-Labs/graphify

Links (1)

Notes on AI Coding Skills & Plugins by Riskhikesh V. (www.linkedin.com)

31 Jul

Rahul Parwal

ai-chapter

ai-skills