ODSP2203 Pre-recorded On demand
Transform your SDLC into a self-improving software factory
Sid Choudhury, Puneet Saraswat
Summary
From Microsoft's AI summary
Autoheal founders Sid Chaudhry and Puneet Saraswat described moving the SDLC to an AI software factory. They cited an Atlassian study saying 50% of engineers lose over 10 hours weekly to coordination, and estimated $12.5 million in lost capacity for 1,000 engineers. The demo showed 100+ integrations, governance policies, audit logging, agents turning a Jira ticket into a merge-ready pull request, YAML definitions, evaluation pipelines and a context graph. Nomura's incident recovery was cited as falling from over two hours to 15 minutes.
Session description
Microsoft's catalogue description
Building AI agents is easy. Keeping them accurate, cost-efficient, governed, reliable, and secure in production is the hard part. As AI multiplies code volume, engineering teams face "agent sprawl" and technical debt. Learn how to transform your SDLC into a self-improving software factory, with real-world examples for private agent evaluations, governance, and cost management, regardless of the coding agent harness.