Thank You, Jensen
Dear Jensen, As I write this, I realize you’ll probably never see it. Still, I wanted to put my thanks into words. I joined NVIDIA on December 15, 2025, almost a year ago, and I write with a heavy heart. You spent more than thirty years building this company and inspired so many of us along the way. You deserved a better ending. To me, NVIDIA will not be the same without you....
Networking and LLM — Supercharge LLM With Domain Expert Knowledge
System Instructions are more potent than you might realize I have seen many posts from Network Engineers and Architects dismissing LLM’s ability to perform domain-specific tasks. While it is true that a generic LLM software like ChatGPT does not possess the same level of domain knowledge, it has a high potential to outperform humans if given proper guidance and some teaching. In this post, I will cut through the noise and examine how to “teach” chatGP to answer domain-specific questions like an expert....
Networking and LLM in the Age of AI — Pt. IV: Knowledge Retrieval using RAG
Using RAG to Enforce ACI Policy Standards Networking management can be complex and depends on accurate information to take necessary actions. Taking the wrong action can be costly. Photo by Susan Q Yin on Unsplash This article is part of a series. Here are the links to the previous articles Part I — Introduction Part II — Exploring Basic Interactions Part III — In-Depth Analysis of ChatGPT’s Responses In this article, we’ll first briefly discuss the basics of RAG, the process, and the components that makeup RAG....
Networking and LLM in the Age of AI — Pt. III: In-Depth Analysis of ChatGPT’s Responses
Analyze LLM responses and understand hallucinations. Efficient Network Operation is about getting precise information without additional “digging in.” This article is a continuation of a series following parts I and II. We will begin by analyzing several complex conversations and conclude with examples illustrating hallucinations. Recap The previous articles covered the basics of instructions, prompts, context, function calls, and how all these components integrate. We also reviewed the responses to a fabric health query....
Networking and LLM in the Age of AI — Pt. II: Exploring Basic Interactions
Instructions, Context, Function Calling and Prompts The most crucial information we must provide to any LLM chat model are the context, detailed instructions, and precise prompts. This article is part of a series — You can find Part I here Instructions Instructions play a crucial role in shaping an LLM into a valuable assistant. Their purpose is to set clear expectations and guide narratives and context, thereby reducing inaccuracies and ensuring more precise responses....
Networking and LLM in the Age of AI — Pt. I: The Exploration Begins
A series to explore how to leverage LLM for networking tasks The rise of chatGPT and large language models (LLM) has opened up the possibility of transforming various domains. However, networking remains largely unexplored. In this series, we’ll explore the potential of LLMs to empower Networking Operations. No AI/ML knowledge is required. For the Busy People - TL;DR For the Series LLM demonstrated significant promise in assisting with IT and networking configuration management and troubleshooting tasks....