Introduction​

Modern large language models face a fundamental limitation: context windows. While frontier models now reach 1 million tokens (Nova Premier, Claude Sonnet 4.5), workloads analyzing entire codebases, document collections, or multi-hour conversations can easily exceed 10 million tokens—far beyond any single model's capacity.

This post demonstrates Recursive Language Models (RLMs), an inference strategy from MIT CSAIL research that enables scaling to inputs far beyond context windows. What makes this implementation special: Strands Agents and Amazon Bedrock AgentCore reduce what could be weeks of glue code and deployment work to just a few hours of development.

Every year, AWS re:Invent brings together thousands of builders, leaders, and innovators to explore the future of cloud. In 2025, the catalog is bigger than ever — 1,843 sessions across 53 areas of interest and 19 industries. Inspiring, yes — but also overwhelming.

That's why I built the re:Invent 2025 MCP Server: a comprehensive Model Context Protocol server that transforms how professionals navigate AWS's flagship conference, providing intelligent access to the complete session catalog with advanced search capabilities and speaker discovery.

Note: This is a curiosity-driven experiment, not a production recommendation. For real workloads, Amazon SageMaker is the right choice. This project explores what's possible when you push serverless boundaries.

1. The idea​

After my BitNet Lambda experiment, I kept thinking: what about embeddings? I had text generation working on Lambda, but what about the other half of modern AI applications?

Google's EmbeddingGemma caught my attention—300M parameters, multilingual, designed for efficiency. Could it work on Lambda? Only one way to find out.

So I fired up Amazon Q Developer and started experimenting.

✨ What you'll learn (tl;dr) In ~12 minutes you'll see how to deploy Microsoft's BitNet 1.58-bit quantized LLM on AWS Lambda, the container-based architecture, and performance benchmarks across different memory configurations using the microsoft/bitnet-b1.58-2B-4T model.

Big idea: 1.58-bit quantization enables LLM deployment on Lambda's CPU infrastructure. At ~1.1GB, the model fits within Lambda's constraints for serverless AI inference.

As large language models (LLMs) evolve from passive responders into autonomous agents that can reason, plan, and act—welcome to the age of Agentic AI. These systems don't just generate answers; they browse the web, execute scripts, send emails, and even orchestrate other agents. And with that autonomy comes an entirely new class of cybersecurity threats.

The OWASP Agentic AI: Threats and Mitigations report is the first of its kind to lay out a structured threat model tailored to the unique risks introduced by LLM-powered agents. From memory poisoning and cascading hallucinations to identity spoofing and rogue agents—this is the new frontline of AI security.

✨ What you'll learn (tl;dr) In ~8 minutes you'll see why most data‑governance efforts stall, how to turn governance into load‑bearing scaffolding, and the exact roadmap, roles, and rituals that move you from ad‑hoc chaos to self‑sustaining confidence—without freezing delivery.

Big idea: Data governance isn't red tape; it's the scaffolding that lets strategic initiatives—from AI to customer experience—scale safely and evolve fast. Bake lightweight governance into culture, rituals, and engineering workflows so raw data turns into durable business value without slowing delivery.