Volume 1 rebuilt the mental model of modern LLM research — transformers, scaling laws, RLHF, reasoning, the 2026 frontier. But there’s a different skill that’s been eating the field this year: actu...
Interactive Field Guide to Modern LLMs: From Transformers to the Agentic Frontier
The pace of LLM research is relentless. If you know deep learning fundamentals but have been heads-down on other things for a year or two, the modern landscape can feel like alphabet soup — GRPO, R...
Unleash the Power of Bare Git Repositories to Manage Your Dotfiles
Dotfiles: the unsung heroes of our development environments, silently configuring our tools and making our lives easier. But what happens when your trusty dotfiles are scattered across multiple mac...
My Summer Internship Experience at Meta: Building Video Understanding Systems and Exploring California (Summer 2022)
Meta Platforms, Inc. is a leading tech company that provides users with a wide range of products and services worldwide. Some of the popular mentions are Facebook, Instagram, and Whatsapp. During t...
Setting up Franka Emika Panda Arm with ROS (Noetic)
The Franka Emika Panda Arm is a seven-axis industrial robot arm that is becoming increasingly popular in robotics research and manufacturing. It is easy to use, has a high payload capacity, and com...
A Beginner's Guide to Semi-Supervised Learning
When we try to divide machine learning methods, we generally tend to separate them into two fundamentally different groups: supervised and unsupervised. So, where do semi-supervised methods fit in?...
Enhance your CNN Networks with Squeeze and Excitation (SE) blocks: Attention Mechanism for Input Channels
We know that Convolutional Neural Networks (CNNs) are very popular when it comes to capturing specific properties of an image that are most salient for a given task. In fact, most of the research h...
An Overview of Research in Video Description in 2020
The following post is greatly inspired by the paper Video Description: A Survey of Methods, Datasets and Evaluation Metrics. I have tried to summarize the research that has been undergone in the fi...
Hyperparameter Tuning in Deep Neural Networks
One way to make your deep learning model more accurate and generate better results is to tune your model’s hyperparameters. By doing so, you can speed up your training process and optimize the outp...
Optimization Algorithms in Deep Learning
Building deep neural networks is one thing, but optimizing it to train faster with better accuracy is a completely different set of domain. So, it is very important that we focus on optimizing our ...