How GitHub Copilot Serves 400 Million Completion Requests a Day
At QCon San Francisco Dave Cheney gave an insightful talk about how GitHub Copilot works under the hood and is able to serve 400 Million completion requests a day.
Cool URIs don't change
Cool URIs don't change, but a domain can change.
A New Home for Home Assistant
Moving My Home Assistant to a new server.
Overview of Messaging in Distributed Systems
This post takes a brief look at the most important concepts of messaging in distributed systems.
Implementing Retrieval-augmented generation (RAG) with an own LLM
Implementing RAG for an own LLM.
Retrieval-Augmented Generation in Large Language Models: The Theory
Retrieval-Augmented Generation can enhance an LLM with a specific knowledge base. This post takes a look at some theory of this topic.
Fine-tuning LLMs Locally vs. In The Cloud
An approach to show the differences between training an LLM on local hardware and in the cloud.
Book review: Building an LLM (From Scratch)
A short review of the LLM from Scratch book by Sebastian Raschka