Experiments at work

In my previous article, WebRTC - What the heck? , we met two shopkeeper farmers, John and Finch, who trade goods across a river using detailed letters (SDPs). If you haven't read it, I'd strongly recommend you read that first. I promised we'd get into the technicals. Here it is. But don't worry, no code. We're sticking with John and Finch to understand what really happens under the hood. The postal service - Signaling server In the first article, Finch sent his letter via a pigeon. But John and Finch can't just throw letters across the river and hope it lands. Someone has to carry the letter. That's the postal service. It doesn't read the letter. It doesn't care what's inside. It just delivers. In WebRTC, this is the Signaling Server . It carries the SDP offer from one client to the other and brings back the SDP answer. It could be a WebSocket server, an HTTP server, or anything that can pass a message between two endpoints. WebRTC doesn...

A while ago, we were working on an SDK, and some third-party developers who used our SDK reached out with an interesting request. We used to maintain extensive documentation in our GitHub Wiki, and they wanted to stay updated every time something changed in it. They didn't want to manually check the wiki every day, and GitHub doesn't send email notifications for wiki changes.  This got me digging around, and that's when I discovered something pretty cool - GitHub has Atom feeds for almost everything! Commits, releases, tags, and yes, even wiki changes. I tried it out myself and suggested the same to them. Problem solved, and I thought, why not share this little gem with everyone? /**  * @disclaimer  * This post covers GitHub's publicly available Atom feeds  * These feeds work for public repositories  * For private repos, you'll need authentication  * Please feel free to leave comments if I've missed something */ What are Atom Feeds? Atom feeds are XML-based...

So you want to build an AI assistant? Maybe for customer support, maybe to search through documents, or maybe just to automate some boring stuff at work. You keep hearing words like "prompting", "RAG", "fine-tuning" thrown around in meetings. But what do they actually mean and which one do you need? Let me walk you through building a customer support AI assistant and show you how each of these concepts fits in. /**  * @disclaimer  * This is an educational post based on my learnings.  * I'll share actual experiments in future posts.  * Your use case might be different, so take what applies to you. */ What even is an LLM? Before we jump into the techniques, let's start with the basics. What is an LLM? Think of a Large Language Model (LLM) as someone who has read most of the internet but has zero memory of your conversation once it ends. Models like GPT-4, Claude, or Llama have been trained on massive amounts of text data. They know a lot...