dardanvuc1996@gmail.com

Large context windows have dramatically increased how much information modern language models can process with a single command. With models capable of handling hundreds of thousands—or even millions—of tokens, it’s easy to imagine that Retrieval-Augmented Generation (RAG) is no longer needed. If you can install an entire codebase or script library in a context window, … Read more

A year ago, AI devs relied on the ReAct (Consult + Act) pattern—a simple loop where the LLM thinks, chooses a tool, and implements. But as any software engineer who’s tried to deploy these agents in production knows, simple loops are tough. They miss things, lose track of complex goals, and struggle with ‘tool noise’ … Read more

In the world of Generative AI, latency is the biggest immersion killer. Until recently, building a voice-enabled AI agent felt like putting together a Rube Goldberg machine: you would input audio into a Speech-to-Text (STT) model, send the transcription to a Large-Language Model (LLM), and finally move the text to a Text-to-Speech (TTS) engine. Each … Read more

In this lesson, we build advanced Griptape-a customer support automation system that integrates decision-making tools and agency thinking to process real-world support tickets in the end. We build custom tools to clean sensitive information, categorize issues, prioritize with clear SLA targets, and generate structured incremental payments, all before engaging the language model. We then use … Read more