Blog | shubham's notes

All Writing

RecSys Dec 18, 2025 - 12 min

Why MAP and MRR are not a good choice for Search Ranking

Search ranking practitioners often use Mean Reciprocal Rank (MRR) and Mean Average Precision (MAP) to assess the quality of their rankings. In this post, we will discuss why MAP and MRR are bad for search ranking. We then look at two metrics that serve as...

LLMs Jul 13, 2025 - 7 min

Multi-Head vs Multi-Query vs Grouped Query Attention

The title may be long, but the explanation isn't. Since the original Transformer, researchers have continuously tweaked its architecture, aiming to cut inference costs. Today, we’ll compare Multi-Head vs Multi-Query vs Grouped Query Attention to understand...

Model Deep Dives Apr 5, 2025 - 13 min

CatBoost- Inner Workings and Optimizations

Gradient boosting is a cornerstone technique for modeling tabular data due to its speed and simplicity. It delivers great results without any fuss. When you look around you'll see multiple options like LightGBM, XGBoost, etc. CatBoost is one such variant. In...

Statistics Mar 22, 2025 - 11 min

Statistical Testing 101

In the era of LLMs, AI agents, and other groundbreaking technologies, one essential concept holds its ground—statistical testing. It serves as the foundation for many companies when evaluating whether a new feature or model resonates with their users. In this...

Model Deep Dives Mar 21, 2025 - 5 min

SGD Classifier vs Logistic Regression

Among all the classifiers provided by Sklearn, two stand out for their similarities: SGDClassifier and LogisticRegression. So, what differentiates the two? In this post, we will explore the key differences and compare SGD Classifier vs Logistic Regression....

LLMs Mar 8, 2025 - 6 min

What is KV-Cache?

If you have been reading articles on LLMs, you would have often come across an interesting term called KV-Cache and how developers are trying to do all sorts of trickery to speed up LLMs. And that is what we are going to do today– talk about KV-Cache to...

LLMs Mar 2, 2025 - 11 min

Fine-Tune Embeddings for RAG to Improve Retrieval

Does your RAG application throw up bizarre, irrelevant content? The issue might not be your LLM but your retrieval. Fine-tune embeddings for RAG can significantly improve retrieval quality. When your RAG application doesn't work, you don't need to throw a...

Python Feb 26, 2025 - 5 min

How I Optimize Memory Usage in Pandas

This is not a blog post about code optimizations, data structures, etc. to reduce memory usage. Although that would be pretty useful too, this isn't that. This is something more fundamental than that– How to Optimize Memory Usage in Pandas! So let's see how...

ML Fundamentals Jul 19, 2024 - 7 min

Huber Loss

In machine learning, handling outliers effectively is crucial for building robust models. Traditional loss functions like Mean Squared Error (MSE) can be overly sensitive to outliers, while Mean Absolute Error (MAE) may not provide smooth gradients for...

RecSys Jul 12, 2024 - 5 min

Multi Armed Bandits

Multi Armed Bandits: the name comes from casino slot machines bandits the idea is to maximize the reward or in other words minimize the regret if you knew whats the best machine you would go ahead and exploit it to get the best reward. But in order to do that...