Running Local AI with AWS as a Control Plane

Most AI discussions assume everything runs in the cloud.In practice, that doesn’t always make sense.

Local models offer privacy, lower cost, and control over your own data. Cloud models offer stronger capabilities, larger context windows, and managed infrastructure. The challenge is figuring out how to use both without turning your system into something overly complex.

In this talk, Art Ianuzzi will explore a simple approach: treating AWS as a control plane for a system that can use local AI, cloud AI, or both depending on the task.

We’ll walk through a minimal, working architecture that routes requests between a local model and a cloud LLM based on a small set of practical decisions. The goal is not to cover every service or build a perfect system, but to cover a basic blueprint that you can extend.

Topics will include:

• When local models make sense vs when to use cloud models
• Simple strategies for routing requests between systems
• How AWS can coordinate, log, and extend a hybrid AI setup
• Tradeoffs around cost, privacy, latency, and capability

• Basic familiarity with AI/LLMs is helpful
• No prior AWS experience required

• Laptop optional
• Questions encouraged

Beginner to Intermediate

AI, LLM, AWS, Machine Learning, Systems Design, Hackers Guild

Know what’s Happening Next — before everyone else does.

Get App

Host or Publisher Hackers Guild PGH