/posts/Recent content in Posts on Chamil JayHugoen-usSat, 15 Aug 2026 00:00:00 +0000/posts/language_models/Sat, 15 Aug 2026 00:00:00 +0000/posts/language_models/<p>Less than five years ago, it would have been difficult to imagine that one of the most impactful and transformative advances in artificial intelligence would come from what, on the surface, looks like one of the simplest forms of AI, that deals with one of the most simplest form of data: <strong>text</strong>.</p> <p>At the time, much of the excitement in AI was centred around increasingly sophisticated models for computer vision, robotics, reinforcement learning, speech, and other specialised domains. Computer vision systems were becoming remarkably good at recognising objects, generating images, understanding scenes, and even creating entirely new visual content.</p>/posts/deep_research_agent/Sun, 24 May 2026 00:00:00 +1000/posts/deep_research_agent/<p>A single search-and-summarize pass can answer straightforward questions, but it tends to break down when the task requires breadth, evidence coverage, and the ability to recognise what is still missing. Deep research is less about making one particularly clever model call and more about designing a workflow that can <strong>plan, gather evidence, assess that evidence, and deliberately search again when necessary</strong>.</p> <p>In this tutorial, we will build a simple <strong>deep research agent pattern</strong> with LangGraph that plans, searches in parallel, reflects on evidence quality, loops when needed, and only then writes a final report. The architecture that I will talk about here is very similar to the architecture that I have used within the research team in my personal project <a href="/personal_projects/finai/">FinAI</a>.</p>/posts/llm-primer/Sun, 01 Mar 2026 00:00:00 +0000/posts/llm-primer/<p><em>Disclaimer: This tutorial is primarily based on a knowledge-sharing session I conducted at work. While the session provided the inspiration and foundation for the content, all examples, explanations, code, and opinions presented here are my own and are intended solely for educational purposes. Nothing in this tutorial represents, reflects, or should be interpreted as the views, practices, or intellectual property of my employer.</em></p> <p>Modern language models can appear remarkably intelligent, but at their core they perform a deceptively simple operation:</p>/posts/pytorchprimer/Sun, 01 Feb 2026 00:00:00 +0000/posts/pytorchprimer/<p>PyTorch is one of the most widely used frameworks for building and training neural networks. At the centre of PyTorch is the <strong>tensor</strong>.</p> <p>A tensor is a general-purpose numerical data structure that can represent anything from a single number to a high-dimensional collection of values. Neural-network inputs, model weights, activations, gradients, images, audio representations, and many other objects are represented using tensors.</p> <p>If you are using an Apple Silicon Mac, PyTorch can also take advantage of Apple&rsquo;s GPU through the <strong>Metal Performance Shaders (MPS)</strong> backend.</p>