🎥 Recorded live at the MLOps World | GenAI Summit 2025 — Austin, TX (October 8, 2025) Session Title: Why CI/CD Fails for AI, and How CC/CD Fixes It Speakers: • Aishwarya Naresh Reganti, Founder, LevelUp Labs (Ex-AWS) • Sai Kiriti, Member of Technical Staff, OpenAI Talk Track: Agents in Production Abstract: Traditional CI/CD pipelines assume deterministic software — but AI systems don’t play by those rules. In this talk, Aishwarya Naresh Reganti (LevelUp Labs, ex-AWS) and Sai Kiriti (OpenAI) introduce the Continuous Calibration / Continuous Development (CC/CD) framework — a new approach for building and scaling AI products that behave unpredictably and autonomously. Drawing on 50+ real-world deployments, they share how teams can start with low-agency, high-control setups, then safely scale AI systems as they earn reliability and trust. Attendees will walk away with a clear blueprint for managing AI’s inherent uncertainty — from eval design to behavioral monitoring and trust calibration. What you’ll learn: • Why CI/CD fails for AI and the risks of applying software assumptions to AI systems • How the CC/CD framework enables continuous calibration and safe autonomy • How to scope capabilities and design meaningful evaluation metrics • How to monitor AI behavior in production and prevent drift • Practical steps to scale autonomy without losing control.

CI/CD Fails for AI & How CC/CD Fixes It