Descriptions:
This recording covers Day 2 of the AI Engineer conference in Singapore, featuring a full-day program of technical sessions from practitioners at Arize AI, LlamaIndex, Ably, Google DeepMind, Cloudflare, and other companies building and deploying AI agents in production. The event brings together engineers, product managers, and applied researchers to share concrete lessons rather than theoretical frameworks.
Among the highlighted sessions, Salian — Head of Product at Arize AI and a core contributor to their agent Alex — walks through four key lessons learned over three years of development: managing agent attention across multi-step tasks, handling context effectively, crystallizing correct agent behavior, and debugging real production failures. Arize’s platform focuses on observability, evaluation, and iterative improvement for AI agents, and the talk draws directly from the team’s experience building Alex 2.0 with reasoning and planning capabilities. Other sessions address internal agent platform design, developer tooling, and the emerging patterns companies use to structure and bless agent-building paths for their engineering teams.
Recurring themes across the day include making agents reliable enough for production, composing multi-agent systems, tightening feedback loops when developers are their own users, and the distinction between building internal tooling versus deploying customer-facing agents. For engineers working on agentic systems, the conference offers a concentrated view of what teams across industries are actively solving in 2026.
📺 Source: AI Engineer · Published May 17, 2026
🏷️ Format: Livestream
