Recently, there’s been a lot of talk about the Model Context Protocol. I started looking into some practical examples of how it’s used and came across a video that really helped clarify things for me. It was a simple example, but it opened up my understanding of how to apply the protocol effectively. Hopefully, it’s useful for you as well. I’m looking forward to implementing something similar myself. It should also help deepen my understanding of how to use these tools more practically—especially for projects that can’t be exposed to the internet but still want to take advantage of LLMs.

I recently watched an interesting interview with John Ousterhout, the author of A Philosophy of Software Design The interview begins by exploring how AI tools are evolving and their potential to transform software engineering. One key takeaway is that AI will likely take over many low-level coding tasks. As a result, the importance of software design is only going to grow. What really stood out to me was Ousterhout’s point that as AI handles more of the implementation details, the responsibility for designing clear, scalable, and maintainable systems will become even more critical for developers.

Andrej Karpathy’s video continues his series on large language models (LLMs), shifting from foundational theory to hands-on usage. Below is a timestamped summary of the key points covered. Introduction (0:00–0:34) Karpathy introduces the video as a follow-up to his prior explanation of how LLMs are trained. This session focuses on how to actually use LLMs in daily life and work. LLM Ecosystem (0:34–2:49) He overviews the current LLM landscape: ChatGPT by OpenAI is described as the “Original Gangster” and most feature-rich.

If you’re diving into LLMs or just curious about how tools like ChatGPT work under the hood, Karpathy’s breakdown is one of the clearest and most accessible resources out there. The following are highlights along with time stamps. Introduction (0:00 - 1:04) What are LLMs? Karpathy begins by explaining what large language models (LLMs) are, highlighting their capabilities and limitations. He also sets the goal of the session: to provide a comprehensive understanding of LLMs for a general audience.

Context So you have a typical project with a backend API and a frontend app built with React or Angular. You want to ensure that every time you push changes to production, they are thoroughly tested and deployed. To achieve this, you set up a CI/CD pipeline that handles the building, testing, and deployment processes automatically. At first, everything works smoothly, but as your project grows, you notice the pipeline is taking longer to complete.

People in the tech industry are often criticized for their lack of “soft skills.” I don’t know about other industries, but working in tech, I find this to be true and I’m not immune to this either. A common example of this are online forums (ex: StackOverflow), where developers may swiftly mark your question for closure or down vote it. You’re lucky if they specify the reason for their decision.

Using real-time communication with SignalR allows users to interact with the application without needing to refresh the screen to get updates. This post provides an overview of setting up SignalR in a .NET Core application. In this example, we’ll use a .NET Core API as the backend and Angular as the frontend. We implemented this type of communication on dashboard pages to display workflow status details. Backend implementation .NET Core API After you include SignalR package in your project, register it during startup.

So, you have database changes—either schema or data changes—that need to be implemented and rolled out across various environments. One approach could be to connect directly to each database server and execute scripts. However, you might not—and indeed, probably should not—have direct read/write access to certain environments like beta or production. Additionally, consider if a client posed the following questions; would you be able to easily respond? What is the state of the database in our Production environment?

This post continues our series on AI and ML. The aim is to enhance understanding of the main types of machine learning, explain the reasons for choosing the appropriate ML approach, and describe the data requirements for each method. Ways to learn There are three main ways of learning in machine learning. Supervised learning Unsupervised learning Reinforcement learning How would one choose an appropriate learning method? Appropriate ML method Selecting the appropriate machine learning method depends on your objectives.

Earlier this week, I participated in a two-day, hands-on introductory course on Machine Learning and AI, which was useful for someone new to the field like myself. The course helped me understand various concepts I had been curious about, including: The relationship between Machine Learning and AI. The definition of a “model” in this context and the process of “training” it. Clarification on terms like “Hugging Face” and “Llama.” The importance of quality data in machine learning projects.