Software release cycles are now too fast for traditional security tools. Rapid iterations and reliance on open-source and cloud-native tech increase vulnerabilities, challenging AppSec teams to keep up. Attackers are taking advantage, targeting applications and exploiting misconfigurations, excessive permissions, and vulnerable plug-ins.

What if your ML system is running perfectly—but making the wrong decisions? This talk explores Ransomware 3.0, where attackers poison models and pipelines instead of locking systems, and shows how AI‑augmented attacks bypass traditional security and how to defend against silent ML compromise. Speaker: Avinish Thakur Software Engineer, miniOrange Pune, India.

Intelligent caching and context compression can cut LLM token costs by about 50% while improving FinOps governance and preventing runaway AI spending. Speaker: Nitesh Pamnani Senior Engineer, miniOrange Pune, India.

Originally recorded in 2025, we look back at how cybercriminals perfected deception during the first half of the year. Now available as an archive recording, the session highlights the second edition of the LevelBlue Threat Trends Report and explores real-world incident data, fast-moving attack chains, and the social engineering techniques that shaped the threat landscape at the time. While the data reflects early 2025, many of the lessons remain relevant for understanding how today’s threat environment evolved.