Skip to main content
Latest
Implementing Parallelized CUDA Programs From Scratch Using CUDA Programming
Engineering Best Practices

Implementing Parallelized CUDA Programs From Scratch Using CUDA Programming

A beginner-friendly guide for curious minds who don't know the internal workings of model.cuda().

· Avi Chawla

8 Fatal (Yet Non-obvious) Pitfalls and Cautionary Measures in Data Science
Engineering Best Practices

8 Fatal (Yet Non-obvious) Pitfalls and Cautionary Measures in Data Science

8 data science lesson I wish I had known earlier.

· Avi Chawla

15 Ways to Optimize Neural Network Training (With Implementation)
Classical ML and Deep Learning

15 Ways to Optimize Neural Network Training (With Implementation)

Techniques that help you become a "machine learning engineer" from a "machine learning model developer."

· Avi Chawla

A Detailed and Beginner-Friendly Introduction to PyTorch Lightning: The Supercharged PyTorch
Classical ML and Deep Learning

A Detailed and Beginner-Friendly Introduction to PyTorch Lightning: The Supercharged PyTorch

Immensely simplify deep learning model building with PyTorch Lightning.

· Avi Chawla

Augmenting LLMs: Fine-Tuning or RAG?
LLM and Fine-tuning

Augmenting LLMs: Fine-Tuning or RAG?

Understanding the tradeoffs between RAG and Fine-tuning, and owning the model vs. using a third-party host.

· Avi Chawla

Implementing DoRA (an Improved LoRA) from Scratch
LLM and Fine-tuning

Implementing DoRA (an Improved LoRA) from Scratch

A powerful LoRA-variant explained in a beginner-friendly way and implemented in PyTorch.

· Avi Chawla

Understanding LoRA-derived Techniques for Optimal LLM Fine-tuning
LLM and Fine-tuning

Understanding LoRA-derived Techniques for Optimal LLM Fine-tuning

LoRA-variants explained in a beginner-friendly way.

· Avi Chawla

Implementing LoRA From Scratch for Fine-tuning LLMs
LLM and Fine-tuning

Implementing LoRA From Scratch for Fine-tuning LLMs

Understanding the challenges of traditional fine-tuning and addressing them with LoRA.

· Avi Chawla

Model Compression: A Critical Step Towards Efficient Machine Learning
Engineering Best Practices

Model Compression: A Critical Step Towards Efficient Machine Learning

Four critical ways to reduce model footprint and inference time.

· Avi Chawla

Deploy, Version Control, and Manage ML Models Right From Your Jupyter Notebook with Modelbit
Engineering Best Practices

Deploy, Version Control, and Manage ML Models Right From Your Jupyter Notebook with Modelbit

Deployment has possibly never been so simple.

· Avi Chawla

PyTorch Models Are Not Deployment-Friendly! Supercharge Them With TorchScript.
Engineering Best Practices

PyTorch Models Are Not Deployment-Friendly! Supercharge Them With TorchScript.

Eliminating the dependence of PyTorch models on Python.

· Avi Chawla

Optimize Your ML Development and Operations with MLflow
Engineering Best Practices

Optimize Your ML Development and Operations with MLflow

The guide that every data scientist must read to manage ML experiments like a pro.

· Avi Chawla

How to Streamline Your Machine Learning Workflow With DVC
Engineering Best Practices

How to Streamline Your Machine Learning Workflow With DVC

The guide that every data scientist must read to manage ML experiments like a pro.

· Avi Chawla

You Cannot Build Large Data Projects Until You Learn Data Version Control!
Engineering Best Practices

You Cannot Build Large Data Projects Until You Learn Data Version Control!

The underappreciated, yet critical, skill that most data scientists overlook.

· Avi Chawla

A Mathematical Deep Dive Into the Curse of Dimensionality
Classical ML and Deep Learning

A Mathematical Deep Dive Into the Curse of Dimensionality

Mathematically understanding the surprising phenomena that arise when dealing with data in high dimensions.

· Avi Chawla

Why Bagging is So Ridiculously Effective At Variance Reduction?
Classical ML and Deep Learning

Why Bagging is So Ridiculously Effective At Variance Reduction?

Diving into the mathematical motivation for using bagging.

· Avi Chawla

Sklearn Models are Not Deployment Friendly! Supercharge Them With Tensor Computations.
Engineering Best Practices

Sklearn Models are Not Deployment Friendly! Supercharge Them With Tensor Computations.

Speed up sklearn model inference up to 50x with GPU support.

· Avi Chawla

Gaussian Mixture Models (GMMs)
Classical ML and Deep Learning

Gaussian Mixture Models (GMMs)

Gaussian Mixture Models: A more robust alternative to KMeans.

· Avi Chawla

HDBSCAN: The Supercharged Version of DBSCAN (An Algorithmic Deep Dive)
Classical ML and Deep Learning

HDBSCAN: The Supercharged Version of DBSCAN (An Algorithmic Deep Dive)

A beginner-friendly introduction to HDBSCAN clustering and how it is superior to DBSCAN clustering.

· Avi Chawla

DBSCAN++: The Faster and Scalable Alternative to DBSCAN Clustering
Classical ML and Deep Learning

DBSCAN++: The Faster and Scalable Alternative to DBSCAN Clustering

Addressing major limitations of the most popular density-based clustering algorithm — DBSCAN.

· Avi Chawla

Formulating and Implementing the t-SNE Algorithm From Scratch
Classical ML and Deep Learning

Formulating and Implementing the t-SNE Algorithm From Scratch

The most extensive visual guide to never forget how t-SNE works.

· Avi Chawla

Bayesian Optimization for Hyperparameter Tuning
Classical ML and Deep Learning

Bayesian Optimization for Hyperparameter Tuning

The caveats of grid search and random search and how Bayesian optimization addresses them.

· Avi Chawla

Formulating and Implementing XGBoost From Scratch
Classical ML and Deep Learning

Formulating and Implementing XGBoost From Scratch

An extensive visual guide to never forget how XGBoost works.

· Avi Chawla

Formulating the Principal Component Analysis (PCA) Algorithm From Scratch
Classical ML and Deep Learning

Formulating the Principal Component Analysis (PCA) Algorithm From Scratch

Approaching PCA as an optimization problem.

· Avi Chawla

You Are Probably Building Inconsistent Classification Models Without Even Realizing
Classical ML and Deep Learning

You Are Probably Building Inconsistent Classification Models Without Even Realizing

The limitations of always using cross-entropy loss in ordinal datasets.

· Avi Chawla

Why R-squared is a Flawed Regression Metric?
Classical ML and Deep Learning

Why R-squared is a Flawed Regression Metric?

The lesser-known limitations of the R-squared metric.

· Avi Chawla

Generalized Linear Models (GLMs): The Supercharged Linear Regression
Classical ML and Deep Learning

Generalized Linear Models (GLMs): The Supercharged Linear Regression

The limitations of linear regression and how GLMs solve them.

· Avi Chawla

Why Sklearn’s Logistic Regression Has no Learning Rate Hyperparameter?
Classical ML and Deep Learning

Why Sklearn’s Logistic Regression Has no Learning Rate Hyperparameter?

What are we missing here?

· Avi Chawla

Why Do We Use log-loss To Train Logistic Regression?
Classical ML and Deep Learning

Why Do We Use log-loss To Train Logistic Regression?

The origin of log-loss.

· Avi Chawla

Why Do We Use Sigmoid in Logistic Regression?
Classical ML and Deep Learning

Why Do We Use Sigmoid in Logistic Regression?

The origin of the Sigmoid function and a guide on modeling classification datasets.

· Avi Chawla