AI Safety & Reliability

​The most dangerous AI isn't the one that fails — it's the one you can't tell is failing. As intelligent systems move into production, the conversation is shifting from capability to control. How do you design agents that reason transparently, fail gracefully, and stay stable across complex workflows? Join us for a conversation amongst builders working on systems that are careful by design and dependable in production. ⚙️ ​Panelists Hailey Nguyen — AI Research @ Worktrace AI Arjun Subedi — CPO @ Astha Prof. PK Prasanna Kumar — Chief AI Trainer @ Business Optima Moderated by: Akash Saraf — CEO & Founder @ TheAgentic Sponsor About TheAgentic: TheAgentic provides complete infrastructure for production-ready AI agents, currently powering companies in biotech, healthtech, cybersecurity, and fintech where accuracy isn't optional. TheAgentic’s Sanscritic is a universal reasoning layer for AI, separating how systems think from the models they run on. Instead of relying on fragile prompts or locked-in model behavior, Sanscritic lets you design, control, and own your reasoning logic across…