Skip to content
HG/os

AI Lab

Agents, prompt engineering, and LLM efficiency

Grounded in production experience building AI agents on AWS Bedrock. Notes are published on LinkedIn.

Article · LinkedIn

The Next AI Breakthrough Might Not Be a Bigger Model

The next leap in AI may come from efficiency, not scale — Google's TurboQuant shows aggressive vector compression can hold model quality while slashing memory and infra cost.

aiefficiencyquantizationrag

Coming soon

LangGraphCrewAIMCP ServersRAG experimentsEvaluation pipelines