Agentic AI, LLM engineering, and the models behind modern automation — multi-agent systems, LLM evaluation and comparisons, RAG, fine-tuning, AI infrastructure, security, and production AI engineering.
From the blog
Deep dive into HandoffInputData, input_filter functions, built-in handoff filters, custom filtering strategies, and input_type with Pydantic models for structured handoff metadata in the OpenAI Agents SDK.
Master the art of debugging multi-agent systems using OpenAI's built-in tracing infrastructure to trace handoffs, profile tool calls, and identify bottlenecks in complex agent pipelines.
A comprehensive guide to deploying OpenAI Agents SDK applications to production using Docker, Kubernetes, environment variable management, health checks, autoscaling, and load balancing.
Build a real-time chat agent using WebSocket connections with FastAPI, OpenAI's streaming responses, persistent conversation state, and response chaining via previous_response_id.
Build long-running AI agents that survive crashes, handle human-in-the-loop approvals, and manage multi-hour workflows using Temporal, DBOS, and Restate with the OpenAI Agents SDK.
Implement encrypted agent sessions for HIPAA and SOC2 compliance using the OpenAI Agents SDK EncryptedSession wrapper with AES-GCM encryption, key management, and TTL expiry.
Learn how to build multi-layer compliance guardrails for enterprise AI agent systems, including audit trails via tracing, data residency enforcement, role-based access control, and regulatory reporting.
Ai agent platform export traces to data warehouse: learn how to use add_trace_processor() to export OpenAI Agents SDK traces to Langfuse, Weights and Biases, Arize, and custom observability platforms with production-ready exporter patterns.
Explore the emerging patterns shaping the future of agentic AI, from agent-to-agent communication protocols and autonomous ecosystems to multi-modal agents, evaluation standards, and trust architectures.
