I recently gave a talk on the shift from manual coding to Agentic Engineering. The core of my argument is simple: If you can model a problem, AI can learn it and solve it. Engineering is just the first industry to witness this transition, but it is coming for all knowledge workers. We are moving fr… read more

Originally drafted in November 2021. Updated with a 2026 lens. Shared data platforms are often proposed as a universal solution: centralize ingestion, storage, query patterns, and governance so product teams can move faster. The promise is attractive. The failure mode is also predictable. This post … read more

For the last decade, our industry mantra has been "Serverless First." We abstracted everything away to Managed Services to move fast, ship features, and let someone else handle the DevOps. That strategy worked when money was free. It doesn't work when margins are shrinking. We are now seei… read more

I'm excited to share another collaboration with AWS on a recent blog post: Build trustworthy AI agents with Amazon Bedrock AgentCore Observability. A key challenge when building and scaling AI agents is the complexity of setting up and managing the underlying infrastructure for observability. Tradit… read more

A recent AWS Machine Learning blog post co-authored by Eitan Sela, a Generative AI and Machine Learning Specialist at AWS, highlights how Nielsen Sports achieved a remarkable 75% cost reduction in their video analysis workflows. They modernized their system, which runs thousands of different machine… read more

Geoffrey Hinton, often called the "Godfather of AI," has been sounding the alarm on the existential risks of the very technology he helped pioneer. For many in the industry, it's easy to dismiss this as alarmist sci-fi. But as architects building these systems daily, we see the micro-scale… read more