Which engineering practice most helps build impactful AI systems?

Carly Taylor

VP, Engineering (Concepts Lab) · Crunchyroll

Focus on fundamentals

Carly Taylor says sticking to fundamentals is what she'd focus on, because some things—like keeping data clean and understanding stakeholders' problems—will never change.

The answer starts at 6:09.

Prabhav Jain

CEO · 11x.Ai

Focus on one core problem

Prabhav Jain says they didn't always lead with focusing on one core problem, because with AI you can build so much that your product becomes a Frankenstein and buyers get confused about what you uniquely do.

Angie Jones

VP Engineering, AI Tools & Enablement · Block

Systems thinking

Angie Jones says systems thinking is what's needed now, because that's the skill and level we're operating at—building the system around which agents operate.

Ang Li

CEO & Co-Founder · Simular

Codifying tasks into code

Ang Li argues tasks should be codified into code, because deterministic tasks should be stored in code; code is deterministic and minimizes the accumulated error risk of long decision chains.

Daniel Herrington

Chief AI Officer · The Zebra

Consistent skills repository

Daniel Herrington says their consistent "skills of skills" repo has been most successful, because it pulls all skills files from git so developers stay consistent and graders are more reliable at flagging reviews.

👨🏾‍💻 Ben Ilegbodu

Senior Software Engineer, TV Client Foundations · Netflix

Deciding what to build

Ben Ilegbodu says deciding what to build has helped most, calling planning and decision making the most durable skill because agents and models aren't the ones talking to customers.