Recent Blog Posts
Insights on AI Agents, Modern Web, and the Future of Engineering.
Building an Agent Risk Register from the OWASP LLM Top 10
How to build a practical agent risk register using the OWASP LLM Top 10 and Agentic Top 10 — with risk scoring, mitigation strategies, and code examples.
Structured Logging for Agent Decisions: Events, Inputs, and Outputs
How to implement structured logging for AI agent systems using OpenTelemetry, AgentTrace, and production observability tools.
Interruptible Agents: Mid-Task Corrections from Users
Patterns for building interruptible agents that support mid-task corrections — state preservation, approval gates, and human-in-the-loop UX best practices.
Hybrid Search for Agent Context: Dense, Sparse, and Keyword
Why hybrid search combining dense vector retrieval and sparse BM25 is essential for agent context, with a three-stage pipeline architecture and code examples.
LLM-as-a-Judge for Agent Outputs: Strengths and Pitfalls
How to use LLMs as evaluators for agent outputs — the strengths, systematic biases, and practical debiasing strategies for production evaluation.
Handling Barge-In: Turn-Taking and Interruption for Voice Agents
How to implement barge-in, turn-taking, and interruption handling in voice AI agents — with VAD architectures, Gemini Live API patterns, and production benchmarks.
Handoff Design: Passing State and Control Between Agents
A deep dive into agent handoff patterns — how to pass state and control between agents cleanly, with SDK implementations and failure mode analysis.
MCP Concepts Explained: Tools, Resources, and Prompts
A developer-friendly explainer of the Model Context Protocol — tools, resources, prompts, and how MCP connects AI applications to external systems.
Toxicity and Safety Filters for LLM Outputs: A Hands-On Guide
A practical guide to implementing toxicity detection and safety filters for LLM outputs, covering guard models, content APIs, and production architectures.
Episodic Memory in Agents: Recording Sessions as Replayable Experiences
Learn how episodic memory enables AI agents to record sessions as replayable experiences, with implementation patterns using LlamaIndex and production architectures.
Google Gemini API Computer Use: Build Agents That See, Reason, and Act
Google brings computer use as a built-in tool in Gemini 3.5 Flash, enabling developers to build AI agents that control browsers, mobile, and desktop environments through screenshots and UI actions.
ChatGPT Astra: What OpenAI GPT-6 Astra Brings to the Table
OpenAI has released its latest flagship model GPT-6 Astra, the most intelligent and aligned AI model to date, setting new benchmarks in computer use, coding, cybersecurity, and scientific research.