The 5 AI Guardrails that Actually Worked

How 4 Engineers Shipped AWS's Fastest Product Launch

Ashish Jha (now Director of Engineering @ Drata) built AWS Audit Manager - one of AWS's fastest-growing services ever - launching in 15 months with 1,000 customers turning it on in the first 24 hours. On this episode of 1 IDEA, Ashish shares how extreme constraints - 3 months to prove it, 8 months to ship - forced the architectural decisions that made it work at scale. We cover: * The $200K/month architecture problem - and the filtering fix that saved it * How 50 customer calls killed features the team thought were critical * Why GDPR broke audit logs - and the authorization layer that fixed it * What happens when 1,000 customers launch in 24 hours vs a planned 3-month ramp * What would be AI-first today - and what stays deterministic in compliance

How Cisco Cut Ship Time by 83%

Manish (VP of Engineering @ Cisco) watched AI write perfect code that created a disaster. When his team asked AI to add traceability, it invented a brand new framework instead of using the one they already had. Multiply that across 1,100 engineers, and suddenly you have dozens of competing implementations of the same thing. For his one idea, Manish breaks down the three phases of AI adoption and exactly how to move from chaos to control.  We cover: * Why AI coding without guardrails will make you lose control of your codebase * The "left shift" that makes AI productive: detailed designs in markdown, not PowerPoints * How his team cut a 12-day feature down to 2 days without sacrificing quality * Why developers now need to become cross-functional architects earlier in their careers Check out Suresh's idea: The Self-Driving Cloud Learn how Sedai can make your cloud autonomous → https://www.sedai.io/ [https://www.sedai.io/]

The Runtime Trick Behind Salesforce's UI Overhaul

Salesforce needed a faster UI runtime, but rewriting everything without forcing customers to migrate seemed impossible. On this episode of 1 IDEA, Kevin Hill (VP of Engineering @ Salesforce) shares how backing his team’s “impossible” ideas led to the one idea that unlocked Salesforce’s UI overhaul. What we cover: * The compatibility-layer idea that let Salesforce run Aura inside the new runtime * How POCs and demos turned critics into allies * How the interop layer removed Salesforce’s migration blocker And, check out Suresh's favorite idea: The Self-Driving Cloud™. - https://www.sedai.io/ [https://www.sedai.io/]

The 5 AI Guardrails that Actually Worked

Rachit Lohani (fmr. CTO @ Paylocity) has seen the same pattern everywhere: engineering leaders getting hammered with "Why aren't we more productive with AI?" Teams are using AI. They're writing more code. But nobody can answer whether it actually matters. For his one idea, Rachit breaks down why AI productivity metrics fail — and the guardrails that make them useful. We cover: * The 5 guardrails to build before building with AI * The two biggest challenges killing real AI adoption — and what to do instead * How to move from vanity metrics to measurable outcomes And, check out Suresh’s favorite idea: The ⁠Self-Driving Cloud™⁠ [https://www.sedai.io/].

The Bottleneck Test

Vinay Perneti [http://linkedin.com/in/vinayperneti], VP of Engineering at Augment, learned the hard way that being the go-to problem solver was actually "diminishing the intelligence" of his team.  Years ago at Pure Storage, his manager's brutal feedback forced him to develop the "Bottleneck Test": a framework he uses to scale teams at Augment in the fast-moving AI space.  For his one idea, he shares how to diagnose if you're accidentally blocking your team and the specific questions that transform dependence into ownership. We cover: * The Bottleneck Test: three diagnostic questions to know if you're accidentally blocking your team * Why being off the critical path isn't passive and how to coach without giving answers * The one-week experiment where you only ask questions and give zero answers * When to jump in and solve a problem yourself and how to do it without undermining your team * Using LLMs as thought partners to reflect on your leadership   Check out Suresh's idea: The Self-Driving Cloud™ Learn how Sedai can make your cloud autonomous → https://www.sedai.io/